RU2794185C2 - Method for determining a patient's response to neuropharmacological intervention, system and machine readable medium for its implementation. - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: group of inventions relates to a method for determining a patient's response to neuropharmacological intervention, a system and a machine-readable medium for its implementation. First structural neurological data are obtained from a patient population before a neuropharmacological intervention. Moreover, these first structural neurological data characterize the physical structure of a plurality of cortical areas. The physical structure is the cortex thickness (CT) and/or bark surface area (SA). The first correlation matrix is created based on the first structural neurological data by assigning a set of structural nodes corresponding to the cortical areas and determining pairwise correlations between structural nodes based at least in part on the corresponding data of the first structural neurological data. Second structural neurological data are obtained from the said patient population after the neuropharmacological intervention. Moreover, these second structural neurological data characterize the said physical structure of a plurality of cortical areas. The second correlation matrix is created based on the second structural neurological data by determining pairwise correlations between structural nodes based at least in part on the corresponding data of the second structural neurological data. EFFECT: group of inventions allows to determine the responses of the tauopathic patient population to neuropharmacological intervention in the form of a drug product with a symptomatic or disease-modifying effect.

20 cl, 45 dwg, 12 tbl

Description

Technical field

The present invention relates to the use of network methods in the study of neurodegenerative diseases.

State of the art

Models of the human brain as a complex network of interconnected subunits have improved our understanding of the normal organization of the brain and made it possible to eliminate functional changes in neurological disorders. These subunits make up the so-called modules of the brain, i.e., groups of areas that have a high density of connections within them, and with a lower density of connections between groups. It has been proposed that the modular organization of the brain underlies the efficient integration between spatially separated neural processes that support various cognitive and behavioral functions. Alterations in brain networks may help identify patients with Alzheimer's disease (AD) and a behavioral variant of frontotemporal dementia (bvFTD).

For example, changes in regional volumes have been identified in patients with schizophrenia through a study of structural networks in health and disease, where pairwise correlation of cortical regional volume or thickness was tested from in vivo T1-weighted magnetic resonance imaging (MRI) measurements. This approach has demonstrated clinical relevance by identifying changes in regional volumes in patients with schizophrenia. However, volumetric measurements, which are the product of bark thickness (CT) and surface area (SA), can confound the underlying differences. For example, looking at changes in cortical thickness can provide insight into how disease changes the size, density, and arrangement of cells in cortical layers. On the other hand, changes in surface area can provide information about impaired functional integration between groups of nerve bundles in a diseased brain.

Previously, the volume of the entire brain or lobes of the brain was used to monitor the effect of neuropharmacological intervention. However, this is a relatively crude level of analysis.

As another example of using networks to understand the brain, as discussed in WO 2017/118733 (whose full content is incorporated herein by this link), EEG data collected from a patient can be used to determine the intensity and direction of electrical flow in the brain .

A poster presentation entitled "Organization of cortical thickness networks in Alzheimer's disease and behavioral variant frontotemporal dementia across brain lobes" was presented by Vuksanovic et al. at the 6th Cambridge Neuroscience Symposium, Neural Networks in Health and Disease September 7-8 2017.

Another poster presentation entitled "Divergent changes in structural correlation networks in Alzheimer's disease and behavioral variant frontotemporal dementia" was presented by Vuksanovic et al. at ARUK Conference 2018, 20-21 March, London, UK.

A presentation entitled "Modular organization of cortical thickness and surface area structural correlation networks in Alzheimer's disease (AD) and behavioral variant frontotemporal dementia (bvFTD)" was given by Vuksanovic, V, at the 10th SINAPSE Annual Scientific Meeting 25 June 2018 Edinburgh.

The essence of the invention

In a first aspect, the invention provides a method for determining a patient's response to a neuropharmacological intervention, comprising the steps of:

obtaining structural neurological data from a plurality of patients prior to a neuropharmacological intervention, wherein said structural neurological data characterizes the physical structure of a plurality of cortical regions;

creating the first correlation matrix based on structural neurological data through

assigning a plurality of structural nodes corresponding to the cortical areas of the brain; And

determining pairwise correlations between pairs of structural nodes based at least in part on corresponding structural neurological data;

obtaining additional structural neurological data from said plurality of patients after the neuropharmacological intervention, wherein said additional structural neurological data characterize the physical structure of a plurality of cortical regions; And

creating a second correlation matrix based on additional structural neurological data by:

determining pairwise correlations between pairs of structural nodes based at least in part on corresponding data from additional structural neurological data;

wherein the method includes:

comparing the first correlation matrix and the second correlation matrix, and thereby determining the patient's response to the neuropharmacological intervention.

Further features of the invention will now be set forth. They are applicable singly or in any combination with any aspect of the invention.

Correlation matrix means the creation of a structural correlation network, which can then be represented by a matrix.

In one embodiment, the patient's response may be in the context of a clinical trial, for example, to evaluate the efficacy of a pharmaceutical in the treatment of a neurocognitive disease. Thus, the patient group (multiple patients) may be a treatment group that has been diagnosed with a disease, or may be a control ("normal") group. Ultimately, the efficacy of a pharmaceutical product can be assessed in whole or in part based on the response of a group of patients determined in accordance with the present invention, optionally compared to a comparison group that did not receive the intervention.

The measured or derived physical structure may be bark thickness and/or surface area. The values for cortical thickness and/or surface area may be averaged values derived from structural neurological data. Structured neurological data can be obtained from magnetic resonance imaging (MRI) or computed tomography data for each patient. Structural neurological data and additional structural neurological data are obtained at different time points. As discussed here, structural neurological data can be obtained using magnetic resonance imaging, computed tomography or positron emission tomography for each patient. These methods per se are well known to those skilled in the art - see, for example, Mangrum, Wells, et al. Duke Review of MRI Principles: Case Review Series E-book. Elsevier Health Sciences, 2018 and "Standardized low-resolution electromagnetic tomography (sLORETA): technical details" Methods Find Exp. Clin. Pharmacol. 2002:24 Suppl. D:5-12; Pascual-Marqui RD etc.

The number of cortical regions may be at least 60, or at least 65. For example, 68. Cortical regions, for example, may be those presented in the Desikan-Killiany Atlas (Desikan et al. 2006).

A p-value may be determined for each pairwise correlation across a plurality of subjects, and may be compared to a significance level, with only p-values less than the significance level being used to generate the respective correlation matrices. When determining pairwise correlation between pairs of structure nodes, the respective values of each structure node can be compared with a reference value and their covariance determined. The level of significance may be referred to as alpha ("α").

Comparing the first correlation matrix and the second correlation matrix may include comparing the number and/or density of inverse correlations in the first correlation matrix with the number and/or density of inverse correlations in the second correlation matrix. In the comparison, node groups corresponding to the same proportion can be identified, and a comparison made between the first and second correlation matrices can use the same proportion.

The assignment of a plurality of structural nodes corresponding to cortical regions of the brain may further include the definition of groups that contain structural nodes corresponding to homologous or non-homologous lobes. Comparing the first correlation matrix and the second correlation matrix may include comparing the number and/or density of correlations between different groups of structural nodes. In other words, comparing the first and second correlation matrices may include comparing correlations between pairs of structural nodes that are not homologous.

In some examples, comparing the first correlation matrix and the second correlation matrix may include comparing the number and/or density of correlations between groups of structural nodes located, respectively, in the frontal lobe (anterior nodes) and the parietal and occipital lobes (posterior nodes). It has been found that in examples of effective neuropharmacological interventions, the number and/or density of inverse correlations between anterior and posterior nodes is reduced. Since inverse correlations are presumably specified as a definition of a compensatory relationship, in which atrophy in one node is associated with hypertrophy in a functionally related node, it should be understood that a decrease in the number and/or density of inverse correlations indicates a decrease in the number of compensatory connections.

Typically, the neurocognitive disease or cognitive disorder is a neurodegenerative disorder that causes dementia, such as tauopathy.

Patients may be diagnosed with a neurocognitive disease, such as Alzheimer's disease or a behavioral variant of frontotemporal dementia. The disease may be moderate or moderate Alzheimer's disease. The disease may be a mild cognitive impairment. However, the discoveries of the authors of the present invention described here are also applicable to other neurocognitive diseases.

Criteria for the diagnosis and treatment of taupathies and other neurocognitive diseases are known in the art and are discussed, for example, in WO2018/019823 and the references cited there.

The disease may be a behavioral variant of frontotemporal dementia (bvFTD). Criteria for diagnosis and treatment of bvFTD are discussed, for example, in WO 2018/041739 and the references cited there.

As explained here, the topology of structural network disruption is different in these two disease states (AD and bvFTD), and both are different from normal aging. Abnormalities are global in nature and are not limited to the frontotemporal and temporoparietal lobes, respectively, in bvFTD and AD, and indicate an increase in both the strength of the global correlation and, in particular, non-homologous interlobar interneuronal connections determined by inverse correlations.

These changes appear to be adaptive in nature, reflecting a coordinated increase in cortical thickness and surface area that compensates for corresponding disturbances in functionally related nodes. The effects were more pronounced in the cortical thickness network in bvFTD and in the surface area network in AD.

The inventors observed that an important change that distinguishes both forms of dementia from the control group of normal elderly people is the emergence of significant inverse correlation networks linking the anterior and posterior regions of the brain, which may be related to functional adaptation or compensation for impairments caused by pathology. In particular, inverse correlations are indicated to indicate the formation of compensatory connections, where atrophy in one node is associated with hypertrophy in a functionally connected node, and it should be understood that a decrease in the number and/or density of inverse correlations indicates a decrease in the number of compensatory connections.

Thus, if the neuropharmacological intervention proves effective, networking is expected to return to that observed in the normal (disease-free) comparison population. If the disease is treated at a sufficiently early stage, networking can be reverted to an almost complete equivalent of normal control. Thus, this method provides an objective means of distinguishing disease-modifying therapies from symptomatic therapies: symptomatic therapies highlight abnormal network architecture and may indeed highlight the risk of transmission of (eg) prion-like disease processes to healthy areas of the brain. Conversely, disease-modifying drugs act in the opposite direction, reducing the need for compensatory input from relatively less affected areas of the brain by normalizing function in areas affected by pathology.

In light of the present focus, it should be understood that analysis of structural or network organization is of particular utility in providing greater power in clinical trials, thereby allowing the use of fewer subjects and shorter treatment times. In particular, in diseases such as mild AD, mild cognitive impairment and pre-moderate cognitive impairment, etc., clinical trial endpoints (cognitive and functional) may be relatively insensitive and therefore require a large number of subjects and/or longer time periods (see WO2009/060191).

Thus, typically, a neuropharmacological intervention will be a pharmaceutical intervention.

Neuropharmacological intervention may be symptomatic treatment. Such compounds include acetylcholinesterase inhibitors (AChEls) - these include tacrine, donepezil, rivastigmine and galantamine. Another symptomatic treatment is memantine. These treatments are described in WO2018/041739.

As explained above, the inventors found an increase in compensatory networks (number and/or density of non-homologous inverse correlations present in patient groups that received such treatment).

The neuropharmacological intervention may be a disease-modifying pharmaceutical rather than symptomatic. These treatments can be distinguished, for example, depending on what happens when the patient stops active treatment. Symptomatic agents delay the symptoms of the disease without interfering with the underlying disease process and do not alter (or at least do not improve) the rate of long-term deterioration after the initial period of treatment. If, after withdrawal, the patient returns to where he would have been without treatment, treatment is considered symptomatic (Cummings, J. L. (2006) Challenges to demonstrating disease-modifying effects in Alzheimer's disease clinical trials. Alzheimer's and Dementia, 2:263-271).

For example, a disease-modifying treatment may be an inhibitor of abnormal protein aggregation, such as a 3,7-diaminophenothiazine compound (DAPTZ). Such compounds are described in WO2018/041739, WO2007/110627 and WO2012/107706. The latter describes leuco-methylthioninium bis(hydromethanesulfonate), also known as leuco-methylthioninium mesylate (LMTM; USAN name: hydromethylthionine mesylate).

The contents of all such WO publications with respect to the DAPTZ compounds they define are specifically incorporated by this cross-reference.

Treatment with LMTM has been shown to reduce compensatory network correlations (especially non-homologous, positive and inverse correlations).

The neuropharmacological intervention may be a disease-modifying pharmaceutical and its effectiveness may be determined by reducing the number and/or density of correlations between the anterior and posterior brain regions of the first correlation matrix and the second correlation matrix.

Thus, it can be concluded that in examples of an effective neuropharmacological intervention (eg, a disease-modifying treatment), the number and/or density of inverse correlations between anterior and posterior nodes is reduced.

The invention can also be used to identify functional adaptation or compensation for impairments caused by pathology in a population of patients, for example, to study "cognitive reserve". The invention may be used in conjunction with conventional diagnostic or prognostic measures. These measures include the Alzheimer's Disease Rating Scale - Cognitive Subscale (ADAS-Cog), National Institute of Neurological and Communicative Disorders - Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA), Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSMIV ) and the Clinical Dementia Rating Scale (CDR).

As explained above, the method for determining patient response to neuropharmacological intervention can, in turn, be used to evaluate different cohorts of patients in clinical trials of neuropharmacological intervention. For example, the method may be to determine the effectiveness of a neuropharmacological intervention in a group of patients. The method can be used to define a group of patients according to their response (eg, in terms of certain correlations/inverse correlations). A group of patients can be identified in relation to their previous use of a neuropharmacological intervention and optionally selected for further treatment according to the patient's response.

***

In a second aspect, the invention provides a method for determining the likelihood of a patient developing one or more neurological disorders, comprising the following steps:

obtaining data characterizing the electrical activity in the patient's brain;

creating a network based at least in part on the received data, where the specified network contains a plurality of nodes and directional connections between nodes, where the network characterizes the flow of electrical activity in the patient's brain;

calculating for each node a difference in the number and/or strength of connections into the node and the number and/or strength of connections out of the node; And

determining, using the calculated differences, the likelihood of the patient developing one or more neurological disorders.

The inventors have shown that even a fairly brief analysis using (eg) brain EEG can be used to potentially identify patients predisposed to one or more neurocognitive diseases (eg, AD). In particular, such individuals (patients or subjects, the terms are used interchangeably) may have a relatively large number of "acceptors" or acceptors that are relatively strong in the posterior lobes and a relatively large number of "donors" or donors that are relatively strong in the temporal lobe. and/or frontal lobes. In apparently normal or prodromal subjects, in preferred embodiments, the method may be more sensitive than commonly used psychometric measures to determine such risk.

***

In the following, optional features of the invention will be set forth. They are applicable singly or in any combination with any aspect of the invention.

The likelihood of a patient developing one or more neurological disorders can be referred to as the patient's predisposition to one or more neurological disorders. The method may include the step of determining a state for each node, where the node is determined to be either an acceptor or a donor based on the calculated difference.

The network may be a network with renormalizable partial directional coherence. Any of the steps of the method can be performed offline, that is, not live on the patient. For example, the acquisition of data may be through network reception of data that has been previously recorded from the patient.

Data indicating electrical activity in the brain may be electroencephalography data. The electroencephalography data may be β-band (rhythm) electroencephalography data. Data indicative of electrical activity in the brain may also be magnetoencephalography data or functional magnetic resonance imaging data.

Determination of a patient's predisposition can be performed using a machine learning classifier. For example, Markov models, support vector machines, random forest or neural networks.

The method may include the step of obtaining a heat map based at least in part on node states, where said heat map indicates the location and/or intensity of nodes identified as acceptors and nodes identified as donors in the patient's brain. This presentation of certain nodes can help (eg ergonomically) in determining the predisposition of the patient.

In determining a patient's predisposition, a comparison can be made between the number and/or intensity of donors in the parietal and/or occipital lobes and the number and/or intensity of acceptors in the frontal and/or temporal lobes. Experimentally, it has been observed that patients predisposed to one or more neurodegenerative diseases (and especially Alzheimer's disease) have relatively high acceptor intensities in the posterior lobes and relatively high donor intensities in the temporal and/or frontal lobes.

The method may further include the step of inferring, using node states, a measure of the degree of left-right asymmetry in the location and/or intensity of nodes in the brain corresponding to acceptors and donors.

The neurological disorder may be a neurocognitive disease, which may be Alzheimer's disease.

A patient's susceptibility to one or more neurological disorders can be determined by comparing the number and/or intensity of acceptor-determined nodes in the posterior lobe with a predetermined value, and/or comparing the number and/or intensity of donor-determined nodes in the temporal and /or frontal lobes with a predefined value. A patient may be at high risk of predisposition if the number and/or intensity of nodes identified as acceptors in the posterior lobe exceeds a predetermined value, and/or if the number and/or intensity of nodes identified as donors in the temporal and/or frontal proportions exceed a predetermined value (eg, based on the "control" subject or subjects identified as low risk, or reference data (eg, historical reference data) obtained from them). In other words, and more generally, the determination of predisposition may be based on whether the patient has more and/or stronger donors and/or acceptors in one area of the brain compared to another area of the brain. For example, if there are more and/or stronger donors in the temporal and/or frontal lobes than expected, and/or if the patient has more and/or stronger acceptors in the posterior lobe than expected based on data from control subjects, the patient may be identified as at risk for a neurodegenerative disorder. Such data from control subjects can be obtained by long-term monitoring after the baseline assessment.

The inventors also observed that the symptomatic treatment(s) increased(s) activity emanating from the frontal lobe compared to the non-drug group.

The method of the invention in accordance with this aspect can be used to assess, test or classify a subject's predisposition to one or more neurological disorders for any purpose. For example, a score or other test result may be used to classify a subject's mental state or illness according to predetermined criteria.

Any person can be a subject. In one embodiment, the subject may be suspected of having a neurocognitive disease or disorder, such as a neurodegenerative or vascular disease, as described herein, or may be a subject not identified as being at risk.

In one embodiment, the method is for early diagnosis or prediction of cognitive impairment, such as neurocognitive disease, in a subject.

The disease may be mild to moderate Alzheimer's disease.

The disease may be mild cognitive impairment.

However, the discoveries of the authors of the present invention, described in this document, are also applicable to other neurocognitive diseases. For example, the disease may be another dementia, such as vascular dementia.

This method can optionally be used to inform further diagnostic steps or interventions for the subject - for example, based on other imaging modalities or invasive or non-invasive biomarker assessments, where such methods are known per se in the art.

In some embodiments, the method may be designed to determine a subject's risk of neurocognitive impairment. Optionally, said risk may be further calculated using additional factors such as age, lifestyle factors, and other measurable physical or mental criteria. This risk may be classified as "high" or "low" or may be presented as a scale or spectrum.

***

From the present description, it will be apparent that in addition to assessing the likelihood of developing one or more neurological disorders, the same methodology can be used to evaluate the effectiveness of a disease-modifying treatment in reducing said risk and/or treating said disease, i. to evaluate the efficacy of a pharmaceutical product in the prevention or treatment of a disease or disorder. This may optionally be in the context of a clinical trial as described herein, eg compared to a placebo, or other normal control.

In particular, the present disclosure indicates that the methods of the invention (eg, based on EEG technology) can provide powerful and sensitive indicators of the impact of a disease on a subject. This opens up the possibility of demonstrating the efficacy of a disease-modifying treatment in smaller groups of subjects (e.g., less than or equal to 200, 150, 100, or 50 in the treatment and comparison groups) and over a shorter interval (e.g., less than or equal to 6, 5, 4 or 3 months) and at earlier disease stages or less severe disease (eg, prodromal AD, MCI, or even pre-MCI) than is possible using currently available methods.

Thus, as discussed above, the method can be used with different cohorts of patients in clinical trials of neuropharmacological intervention, for example, a patient group (multiple patients) can be a treatment group that has been diagnosed with a disease (e.g., an early stage disease) treated with a putative agent. disease-modifying treatment compared with the placebo group.

Thus, in one additional aspect, the method steps of the second aspect are used to determine the status or severity of a disease in a patient, rather than to determine whether the patient is likely to develop one or more neurological disorders. This status, in turn, can be tracked through clinical management or a clinical trial.

Thus, in another aspect of the invention, a method is provided for determining a patient's response to a neuropharmacological intervention against a neurological disorder, comprising the steps preceding the neuropharmacological intervention:

(a) obtaining data characterizing the electrical activity in the patient's brain;

(b) creating a network based at least in part on the received data, where the specified network contains a plurality of nodes and directional connections between nodes, where this network characterizes the flow of electrical activity in the patient's brain;

(c) calculating for each node the difference in the number and/or strength of connections into the node and the number and/or strength of connections out of the node; And

(d) determining, using the calculated differences, the patient's status with respect to a neurological disorder;

(e) repeating steps (a) to (d) after the neuropharmacological intervention to determine the patient's further status with respect to the neurological disorder; And

(f) determining, based on said first status and said second status (eg, by comparing them), the patient's response to the neuropharmacological intervention.

Optionally, steps (e) and (f) are repeated and subsequent statuses are used to determine the patient's response over time.

Thus, these methods of the second and additional aspects (and the corresponding systems discussed below) can be used for both clinical trials and clinical management. From a clinical management perspective, a high degree of certainty (e.g., 70%, 80%, 90%, or 95% chance) that electrical activity in a patient's brain (e.g., as measured by EEG) is abnormal in a "normal" individual ( i.e. currently undiagnosed) may be a strong indication for the immediate initiation of medical treatment for dementia. An EEG can also be used at intervals, such as 1, 2, 3, 4, 5, or 6 months, to monitor response to treatment. Conversely, a person with a lower probability of having an abnormal EEG (eg, 30%, 40%, 50%, 55%, or 60%) can be observed more closely at monthly, bimonthly, or three-monthly intervals. Further tests by other methods appropriate to the disorder that are known in the art (eg, biomarker assessment based on amyloid, tau-PET, or CSF) may optionally be used in conjunction with this method.

The optional features with respect to the methods of the second aspect apply, mutatis mutandis, to this aspect as well.

In a third aspect, the invention provides a system for determining a patient's response to a neuropharmacological intervention, comprising:

data collection means configured to acquire structural neurological data from a plurality of patients prior to neuropharmacological intervention, wherein said structural neurological data characterizes the physical structure of a plurality of cortical regions;

a correlation matrix generator configured to generate a first correlation matrix based on structural neurological data via:

the assignment of a plurality of structural nodes corresponding to the cortical regions of the brain; And

determining pairwise correlations between pairs of structural nodes based at least in part on the respective structural neurological data;

where the data collection means is also configured to obtain additional structural neurological data from the mentioned plurality of patients after neuropharmacological intervention, where these additional structural neurological data characterize the physical structure of the plurality of cortical regions; And

the correlation matrix generator is also configured to generate a second correlation matrix based on additional structural neurological data via:

determining pairwise correlations between pairs of structural nodes based at least in part on corresponding data from additional structural neurological data;

where the system further includes either:

display means for representing the first correlation matrix and the second correlation matrix; or

a comparison means for comparing the first correlation matrix and the second correlation matrix and thereby determining the patient's response to the neuropharmacological intervention.

In the following, optional features of the invention will be set forth. They are applicable singly or in any combination with any aspect of the invention.

The correlation matrix may mean the creation of a structural correlation network, which can then be represented by a matrix.

The measured or derived physical structure may be bark thickness and/or surface area. The values for cortical thickness and/or surface area may be averaged values derived from structural neurological data. Structured neurological data can be obtained from magnetic resonance imaging (MRI) or computed tomography data for each patient. Structural neurological data and additional structural neurological data are obtained at different time points. As discussed here, structural neurological data can be obtained using magnetic resonance imaging, computed tomography or positron emission tomography for each patient. These methods per se are well known to those skilled in the art - see, for example, Mangrum, Wells, et al. Duke Review of MRI Principles: Case Review Series E-book. Elsevier Health Sciences, 2018 and "Standardized low-resolution electromagnetic tomography (sLORETA): technical details" Methods Find Exp. Clin. Pharmacol. 2002:24 Suppl. D:5-12; Pascual-Marqui RD etc.

The number of cortical regions may be at least 60, or at least 65. For example, 68. Cortical regions, for example, may be those presented in the Desikan-Killiany Atlas (Desikan et al. 2006).

The display means may display each of the first correlation matrix and the second correlation matrix on a display, wherein the correlation values in each correlation matrix are given a color corresponding to the relative amplitude or strength of the correlation.

The checker can be configured to determine the p-value for each pairwise correlation and compare the p-value for each pairwise correlation, and can compare the p-value with a significance level, the correlation matrix generator can be configured to use only p-values below the adjusted level significance when creating a correlation matrix. The level of significance may be referred to as alpha ("α").

The comparator may be configured to compare the number and/or density of inverse correlations in the first correlation matrix with the number and/or density of inverse correlations in the second correlation matrix. In the comparison, node groups corresponding to the same proportion can be identified, and a comparison made between the first and second correlation matrices can use the same proportion.

The assignment of a plurality of structural nodes corresponding to cortical regions of the brain may further include the definition of groups that contain structural nodes corresponding to homologous or non-homologous lobes. The comparator may be configured to compare the first correlation matrix and the second correlation matrix by comparing the number and/or density of correlations between different groups of structural nodes. In other words, comparing the first and second correlation matrices may include comparing pairs of structural nodes that are not homologous.

The comparator may be configured to compare the first correlation matrix and the second correlation matrix by comparing the number and/or density of correlations between groups of structural nodes located respectively in the frontal lobe (anterior nodes) and the parietal and occipital lobes (posterior nodes). It has been found that in examples of effective neuropharmacological intervention, the number and/or density of inverse correlations between anterior and posterior nodes is reduced. Since inverse correlations are presumably specified as a definition of a compensatory relationship, in which atrophy in one node is associated with hypertrophy in a functionally related node, it should be understood that a decrease in the number and/or density of inverse correlations indicates a decrease in the number of compensatory connections.

In one embodiment, the patient's response may be in the context of a clinical trial, for example, to evaluate the effectiveness of a pharmaceutical drug in neurocognitive disease. Thus, the patient group (multiple patients) may be a therapeutic group that has been diagnosed with a disease, or possibly a control ("normal") group. Ultimately, the effectiveness of a pharmaceutical preparation can be assessed in whole or in part based on the response of a group of patients, determined in accordance with the present invention.

As explained in connection with the first aspect, a neurocognitive disease is usually a neurodegenerative disorder causing dementia, such as tauopathy.

Patients may be diagnosed with a neurocognitive disease, such as Alzheimer's disease or a behavioral variant of frontotemporal dementia. The disease may be mild or moderate Alzheimer's disease. The disease may be a mild cognitive impairment.

Criteria for the diagnosis and treatment of taupathies and these disorders are discussed, for example, in WO2018/019823 and the references cited there.

The disease may be a behavioral variant of frontotemporal dementia (bvFTD). Criteria for diagnosis and treatment of bvFTD are discussed, for example, in WO 2018/041739 and the references cited there.

As explained here, the topology of structural network disruption is different in these two disease states (AD and bvFTD), and both are different from normal aging. These changes appear to be adaptive in nature, reflecting a coordinated increase in cortical thickness and surface area that compensates for corresponding disturbances in functionally related nodes.

Thus, if the neuropharmacological intervention is effective, the networking is expected to return to normal. If the disease is treated at a sufficiently early stage, networking can be brought back to normal, indicating a halt or reversal of the disease state. Thus, the system provides an objective means of distinguishing disease-modifying therapies from symptomatic treatments, as described above.

Typically, the neuropharmacological intervention is a pharmaceutical intervention.

Neuropharmacological intervention may be symptomatic treatment as described above.

For example, a disease-modifying treatment may be an inhibitor of abnormal protein aggregation, such as a 3,7-diaminophenothiazine compound (DAPTZ), as described above.

In a fourth aspect, the invention provides a system for determining a patient's predisposition to one or more neurological disorders, wherein the system includes:

data collection means configured to obtain data indicative of electrical activity in the patient's brain;

network building means configured to create a network based at least in part on the received data, where said network comprises a plurality of nodes and directional connections between nodes, where the network characterizes the flow of electrical activity in the patient's brain;

difference calculation means configured to calculate, for each node, a difference between the number and/or strength of connections into the node and the number and/or strength of connections out of the node; and either:

a display means configured to display a representation of the calculated differences; or

determining means configured to determine, using the calculated differences, the patient's susceptibility to one or more neurological disorders.

As described above with respect to the second aspect, the inventors have shown that even a fairly brief analysis using (eg) brain EEG can be used to potentially identify patients who are susceptible to one or more neurocognitive diseases (eg, AD).

The system can be used for both clinical trials and clinical management.

Thus, in a further aspect, a system is provided, as described above, for determining a patient's response to a neuropharmacological intervention against a neurological disorder. In this aspect, the determination means system may be configured to determine, using the calculated differences, the patient's status with respect to a neurological disorder.

The system can be used to determine the patient's further status after neuropharmacological intervention and is optionally configured to determine, based on the first and one or more subsequent statuses, the patient's response to neuropharmacological intervention, as described above in relation to the corresponding method.

Other additional features of the invention will now be set forth. They are applicable singly or in any combination with any aspect of the invention.

The system may include a state determiner configured to determine a node as either an acceptor or a donor based on the computed difference.

The system may include a status determiner configured to determine a node as either an acceptor or a donor based on the calculated difference.

The network may be a network with renormalizable partial directional coherence. Any of the steps of the method can be performed offline, that is, not live on the patient. For example, the acquisition of data may be through network reception of data that has been previously recorded from the patient.

Data indicating electrical activity in the brain may be electroencephalography data. The electroencephalographic data may be β-band electroencephalographic data. Data indicative of electrical activity in the brain may also be magnetoencephalography data or functional magnetic resonance imaging data.

The determiner may be configured to apply a machine learning classifier. For example, Markov models, support vector machines, random forest or neural networks.

The display means may be configured to present a heat map indicating the location and/or intensity of nodes identified as acceptors and nodes identified as donors in the brain. This presentation of certain nodes can help (eg ergonomically) in determining the predisposition of the patient.

The determination means may compare the number and/or intensity of donors in the parietal and/or occipital lobes and the number and/or intensity of acceptors in the frontal and/or temporal lobes. It has been observed experimentally that patients predisposed to one or more neurodegenerative diseases (and especially Alzheimer's disease) have a relatively high number and/or intensity of acceptors in the posterior lobes and a relatively high number and/or intensity of donors in the temporal and/or frontal lobes.

The system may further comprise an asymmetry mapper configured to output, using the node states, an indicator of the degree of left-right asymmetry in the location and/or density of nodes in the brain corresponding to acceptors and donors.

The neurological disorder may be a neurocognitive disease, which is not necessarily Alzheimer's disease.

The determining means may compare the number and/or intensity of nodes identified as acceptors in the posterior lobe with a predetermined value, and/or compare the number and/or intensity of nodes identified as donors in the temporal and/or frontal lobes with a predetermined value. The determining means may determine the patient as having a high risk of predisposition if the number and/or intensity of nodes identified as acceptors in the posterior lobe exceeds a predetermined value and/or if the number and/or intensity of nodes identified as donors in the temporal and/or frontal fractions exceeds a predefined value. In other words, and more generally, the determination of predisposition may be based on whether the patient has more and/or stronger donors and/or acceptors in one region of the brain compared to another region of the brain. For example, if there are more and/or stronger donors in the temporal and/or frontal lobes than expected, and/or if the patient has more and/or stronger acceptors in the posterior lobe than expected, then the patient may be defined as at risk for a neurodegenerative disorder.

The inventors also observed that the symptomatic treatment increased activity emanating from the frontal lobe compared to the non-drug group.

The system of the present invention in accordance with this aspect can be used to assess, test or classify a subject's predisposition to one or more neurological disorders for any purpose. For example, a score or other test result may be used to classify a subject's mental state or disease state according to predetermined criteria.

Any person can be a subject. In one embodiment, the subject may be suspected of suffering from a neurocognitive disease or disorder, such as a neurodegenerative or vascular disease as described herein, or possibly a disease that is not at risk.

In one embodiment, the system is for early diagnosis or prediction of cognitive impairment, such as neurocognitive disease, in a subject, as described above.

The system can additionally be used to inform further diagnostic steps or interventions for the subject - for example, based on other systems for imaging or invasive or non-invasive biomarker assessments, where such systems are known as such in the art.

In some embodiments, the system may be designed to determine a subject's risk of neurocognitive impairment. Optionally, said risk may be further calculated using additional factors such as age, lifestyle factors, and other measurable physical or mental criteria. This risk may be classified as "high" or "low" or may be presented as a scale or spectrum.

Like the methods described herein, the system can be used in the context of clinical trials to evaluate the effectiveness of a neuropharmacological intervention. The system can be used to demonstrate the effectiveness of a disease-modifying treatment, such as LMTM, in a relatively small number of subjects (eg, 50) over a relatively short period of time (eg, 6 months) and in the early stages of disease (eg, mild cognitive impairment or possible pre-moderate cognitive impairment).

Additional aspects of the present invention are: a computer program comprising executable code that, when executed on a computer, causes the computer to perform the method of the first or second aspect; a computer-readable medium storing a computer program containing code that, when executed on the computer, causes the computer to perform the method of the first or second aspect; and a computer system programmed to perform the method of the first or second aspect. For example, a computer system may be represented including: one or more processors configured to: execute the method of the first or second aspect. The system thus corresponds to the method of the first or second aspect. The system may further include: a computer-readable medium or media operatively connected to the processors, wherein the medium or media stores computer-executable instructions in accordance with the method of the first or second aspects.

Brief description of the drawings

Embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which:

Figure 1 shows an example of a Desikan-Killiany brain Atlas;

Figure 2 shows an example cortical surface area correlation matrix with pairwise correlations, grouped by lobes, for a group of subjects diagnosed with a behavioral variant of frontotemporal dementia;

Figures 3A-3C show, respectively, group-based cortical thickness correlation networks depicted as pairwise correlation matrices for: (i) the HE subject group, (ii) the bvFTD subject group, and (iii) the AD subject group;

Figures 4A-4C show, respectively, group-based surface area correlation networks depicted as pairwise correlation matrices for: (i) the HE subject group, (ii) the bvFTD subject group, and (iii) the AD subject group;

Figure 5 shows plots of mean edge strength of cortical thickness correlation network averaged across cerebral lobes and compared between HE, bvFTD and AD groups, where the top plot shows networks of positive correlations and the bottom plot shows networks of inverse correlations;

Figure 6 shows plots of mean edge strength of surface area correlation network averaged over brain lobes and compared between HE, bvFTD and AD groups, where the left plot shows networks of inverse correlations and the right plot shows networks of positive correlations;

Figure 7 shows graphs of network node degrees of cortical thickness correlation averaged across the lobes and compared between the HE, bvFTD and AD groups, where the top graph shows networks of positive correlations and the bottom graph shows networks of inverse correlations;

Figure 8 shows graphs of indices of interlobar participation of nodes in the network of correlation of cortical thickness, averaged over the lobes of the brain, and compared between groups HE, bvFTD and AD for positive correlations;

Figure 9 shows plots of network node degrees of surface area correlation averaged over brain lobes and compared between HE, bvFTD and AD groups, where the top plot shows networks of positive correlations and the bottom plot shows networks of inverse correlations;

Figure 10 shows graphs of interlobar node involvement indexes of surface area correlation averaged across brain lobes and compared between HE, bvFTD and AD groups, where the top graph shows networks of positive correlations and the bottom graph shows networks of inverse correlations;

Figure 11 shows brainspace visualization of nuclei in the cortical thickness network for, in the upper plot, positively correlated nodes and, in the lower plot, inversely correlated nodes for the HE, bvFTD, and AD groups;

Figure 12 shows the brain space visualization of the nuclei in the surface area network for, in the upper plot, positively correlated nodes and, in the lower plot, inversely correlated nodes for the HE, bvFTD, and AD groups;

Figure 13 shows a brainspace visualization of the interaction between the positive networks of cortical thickness and cortical surface area for the HE, bvFTD, and AD groups;

Figure 14 shows histograms of preserved edges in cortical thickness (upper three plots) and surface area (lower three plots) correlation networks;

Figure 15 is a graph showing the distribution of the modularity index (Q) in regional cortical thickness correlation networks generated from 100 surrogate datasets;

Figure 16 shows binarized network correlation matrices for bark thickness (upper three panels) and surface area (lower three panels) for the HE, bvFTD, and AD groups, white indicates significant positive correlations and black indicates significant inverse correlations;

Figures 17A-17D show correlation matrices at baseline (i.e. week 01) according to treatment status with symptomatic drugs for AD (cholinesterase inhibitors and/or memantine) where ach0 indicates no treatment and ach1 indicates the presence of such treatment;

Figures 18A-18D show plots of nodal rates of non-homologous interlobar correlations at baseline nodal rates according to treatment status with symptomatic drugs for AD (acetylcholinesterase inhibitors and/or memantine);

Figures 19A and 19B show time-separated cortical thickness correlation matrices at baseline (week 01) and after 65 weeks (week 65) in patients treated with 8 mg/day LMTM in combination with symptomatic treatment;

Figures 20A-20D show plots of positive and inverse non-homologous interlobar nodal correlations in cortical thickness (CT) and surface area (SA) at baseline and after 65 weeks in patients treated with 8 mg/day LMTM in combination with symptomatic treatment;

Figures 21A and 21B show time-separated cortical thickness correlation matrices at baseline (week 01) and after 65 weeks (week 65) in patients receiving 8 mg/day LMTM as monotherapy (i.e. not in combination with symptomatic treatment AD);

Figures 22A and 22B show plots of non-homologous interlobar nodal grades for the cortical thickness correlation network at baseline and 65 weeks later in patients treated with 8 mg/day LMTM in combination as monotherapy (i.e., not in combination with symptomatic AD treatment). );

Figures 23A and 23B show time-separated surface area correlation matrices at baseline (week 01) and after 65 weeks (week 65) in patients receiving 8 mg/day LMTM as monotherapy (i.e., not in combination with symptomatic treatment). AD);

Figures 24A and 24B show plots of non-homologous interlobar nodal rates for the surface area correlation network at baseline and at 65 weeks in patients receiving 8 mg/day LMTM in combination as monotherapy (i.e., not in combination with symptomatic AD treatment). );

Figures 25A-25D show time-separated cortical thickness correlation matrices for both AD (clinical dementia score 0.5, 1 and 2) at baseline and after 65 weeks of treatment with LMTM 8 mg/day as monotherapy and for an elderly control group. humans (HE) to show matrix normalization after treatment;

Figures 26A - 26D show time-separated surface area correlation matrices for both AD (dementia clinical score 0.5, 1 and 2) at baseline and after 65 weeks of treatment with LMTM 8 mg/day as monotherapy and for an elderly control group. humans (HE) to show matrix normalization after treatment;

Figures 27A - 27D show time-separated cortical thickness correlation matrices for both AD (clinical dementia score 0.5) at baseline and after 65 weeks of treatment with LMTM 8 mg/day as monotherapy and for a control group of elderly (HE) to show matrix normalization after treatment;

Figures 28A - 28D show time-separated surface area correlation matrices for both AD (clinical dementia score 0.5) at baseline and after 65 weeks of treatment with LMTM 8 mg/day alone, and for a control group of elderly (HE) to show matrix normalization after treatment;

Figure 29 shows an example of electroencephalography data at rest;

Figure 30 shows an example of a directed network derived from electroencephalography data;

Figure 31 schematically shows the definition of the state of the node as the difference between the incoming flow of electrical activity and the outgoing flow of electrical activity;

Figure 32 shows a heat map of the location of net acceptors (yellow/red) and net donors (blue) in the brain of a group of subjects;

Figure 33 is a heat map showing the asymmetry in the distribution of donors and acceptors between the left and right sides of the heat map in Figure 32;

Figure 34 shows a heat map of the location of donors and acceptors in the brain of a group of subjects diagnosed with Alzheimer's disease;

Figure 35 shows a heat map of the location of donors and acceptors in the brain of a group of subjects who have not been diagnosed with Alzheimer's disease (ie paired volunteers);

Figure 36 shows a heat map of the location of donors and acceptors in the brain of a subject who has not been diagnosed with Alzheimer's disease (ie, a paired volunteer);

Figure 37 shows a heat map of the location of donors and acceptors in the brain of a subject diagnosed with Alzheimer's disease;

Figure 38 shows a heatmap of the location of donors and acceptors in the brain of a group of subjects found to be at risk for dementia or cognitive impairment, for example due to Alzheimer's disease;

Figure 39 shows a heat map of the location of donors and acceptors in the brain of a group of subjects found not to be at risk for Alzheimer's disease;

Figure 40 is a box and whisker plot comparing donors and acceptors in EEG networks from frontal and posterior brain regions at the group level in subjects at risk for AD and not at risk for AD;

Figure 41 shows a comparison of cortical thickness correlation matrices for the AD group, showing the increase in strength and number of significant inverse non-homologous interlobar correlations (left), and a heat map of the location of donors and acceptors in the brain of a group of subjects diagnosed with AD, showing the correspondence between an increase in compensatory structural inverse non-homologous correlations in the thickness of the cortex directed to the posterior regions of the brain, and an increase in the strength and number of incoming connections with the posterior regions of the brain as an acceptor, shown by analysis of EEG rPDC coherence at rest;

Figure 42 shows a comparison between the surface area correlation matrix for the HE group and a heat map of the location of donors and acceptors in the brain of a group of healthy elderly subjects, showing the correspondence between the relative absence of compensatory structural inverse non-homologous correlations in the thickness of the cortex directed to the posterior regions of the brain, and a decrease the strength and number of incoming connections to the posterior brain regions as an acceptor, as shown by resting-state rPDC EEG coherence analysis;

Figure 43 is a box and whisker plot showing the quantitative differentiation of moderate AD from a control group of older adults;

Figure 44 shows three heat maps comparing drug- and non-drug AD patients and paired volunteers at the group level; And

Figure 45 is a box-and-whisker plot of the network at the group level comparing drug- and non-drug AD patients and paired volunteers.

Detailed description and additional optional features

Further aspects and embodiments of the present invention will be discussed with reference to the accompanying figures. Additional aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference.

Figure 1 shows an example of a Desikan-Killiany brain Atlas. The Desikan-Killiany brain Atlas divides the human cortex on MRI scans into regions of interest based on cerebral convolutions. Although the figure shows 18 regions, the complete Desikan-Killiany brain Atlas divides the human cerebral cortex into 68 regions of interest.

The subjects discussed in this paper have participated in three phase 3 global clinical trials that have been completed to date. Two of the trials were for mild to moderate AD (Gauthier et al., 2016; Wilcock et al., 2018) and the third was from a large bvFTD study (Feldman et al., 2016). Comparable data were available from well-characterized healthy elderly (HE) subjects participating in the Aberdeen 1936 Birth Cohort (ABC36) indefinite longitudinal study (Murray et al., 2011). A total of 628 subjects were involved in the examples discussed here, of which 213 were in each of the dementia groups and 202 healthy elderly subjects. Patients with bvFTD were diagnosed according to the International Consensus Criteria for mild bvFTD on the Mini Mental Status Assessment (MMSE) score of 20 to 30 inclusive. Patients with AD were diagnosed according to the criteria of the National Institute of Aging and the Alzheimer's Association, mild to moderate according to an MMSE score of 14-26 (inclusive) and an overall Clinical Dementia Rating Scale (CDR) score of 1 or 2. They were selected from the corresponding larger group (N=1132) to match the number of participants in the bvFTD group. Healthy elderly (HE) subjects were selected from the well-characterized Aberdeen 1936 Birth Cohort.

The imaging datasets from a multilateral source used to create the correlation matrices discussed below were standard T1-weighted MRI images acquired using manufacturer-specified 3DT1 equivalent sequences. Data from patients in the trial were pooled to allow for overall group comparisons. Serial scanners were limited to field strengths of 1.5 T and 3 T (30%) from three manufacturers (Philips, GE and Siemens). All MRI images in the ABC36 cohort were acquired using the same 3T scanner (Philips). The images were processed using an automated process pipeline implemented in a manner known per se. In addition to volumetric imaging modalities, the pipeline produces surface regional measurements of cortical morphology such as thickness, local curvature, or surface area. An example of automated processing pipelines suitable for the above methods is FreeSurfer v5.3.0 available from Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital.

The surface area is calculated from the imaging datasets using a triangular tiling diagram of the grey/white matter interface and the white matter/cerebrospinal fluid interface (referred to as the pial surface). The thickness of the cortex is calculated as the average distance from the surface of the white matter to the nearest point on the pial surface and from this point back to the nearest point on the surface of the white matter. The granulation scheme, known per se, is used to determine the cortical thickness and surface area of 68 cortical regions from both hemispheres based on the Desikan-Killiany Atlas. The list of regions and their distribution by shares is given in Table A.1 in Appendix A.

Figure 2 shows a cortical surface area correlation matrix for a group of subjects diagnosed with bvFTD. Each element of the matrix represents the strength of the correlation ("edge strength") between the 68 pairs of cortical surface areas from the Desikan-Killiany Atlas. The intensity column on the right shows the correlation/strength of the edge. Sixty-eight areas of the cortical surface (network nodes) are arranged in the order of their belonging to the frontal, temporal, parietal, and occipital lobes. Separate areas of the lobes are enclosed in squares and arranged in order from top to bottom / left to right: frontal, temporal, parietal and occipital. In fact, the correlation matrix is a network constructed from partial correlations between 68 pairs of cortical thicknesses. Figures 3A-3C show cortical surface area correlation matrices for, respectively, healthy elderly subjects, subjects with a behavioral variant of frontotemporal dementia, and Alzheimer's disease. Marked differences can be seen between healthy elderly subjects and subjects with bvFTD and AD. Notably, intra-lobal correlation increased significantly in subjects with bvFTD and AD. In addition, the number of inverse correlations increased between non-homologous nodes. As can be seen, HE subjects have sparse correlations, and mostly positive correlations between homologous lobes. At the same time, both bvFTD and AD have a significantly increased number of nodes associated with positive and inverse correlations compared to the HE group. An increase in the number of correlations in both forms of dementia can occur between the same shares (homologous, mostly positive) or between different shares (not homologous, mostly negative). As a rule, inverse interlobar non-homologous correlations are highly anomalous. Furthermore, it can be seen that bvFTD in particular is associated with a higher density of inverse non-homologous correlations across the cortical thickness.

Figures 4A-4C show surface area correlation matrices, respectively, for healthy elderly subjects, subjects with a behavioral variant of frontotemporal dementia, and subjects with Alzheimer's disease. Marked differences can be seen between healthy elderly subjects and subjects with bvFTD and AD. In addition, it should be noted that AD is particularly associated with a higher density of inverse non-homologous correlations in surface area.

In these figures, zero entries correspond to non-significant correlations. Significant network correlations were found to be both positive and negative (see Figures 14 and 16 for illustration). Because of the number of significant inverse correlations and the apparent increase in network correlation strength in bvFTD and AD compared to the HE group, subnets of significant positive and inverse correlations are considered separately. Given the apparent differences in the structure of the proportion network depending on the diagnostic group, an attempt was made to determine whether these differences can be quantified.

The networks, represented in these figures as correlation matrices, can be constructed by correlating either surface area or cortical thickness across all subjects within a particular diagnostic category (ie, HE, bvFTD, and AD). A cortical region (as defined in the Desikan-Killiany brain Atlas) is a node, and a pairwise correlation between nodes, representing a graph edge or connection/connection, was constructed by correlating either SA or CT for all participants in each diagnostic category. Each correlation matrix was calculated based on an S x N matrix containing N regional CT/SA values from S subjects in each group. Thus, six N x N (eg, 68×68) correlation matrices are obtained (one CT or SA structural correlation matrix for each study group). Matrix element e _ij is the partial correlation value between region i and j ( i, j =1, 2 ... N) (ie, between vectors x _i and x _j , which contain regional measurements from subjects in each group). Partial correlations are calculated as linear Pearson correlation coefficients between pairs x _i and x _j after first removing the effects of all other areas m ≠ i, j and then adjusting both x _i and x _j for the control variables (stored in a separate matrix S x C, where C represents the number of control variables). This means that prior to correlation analysis, a linear regression is performed for each x _i to remove the effect of age, sex, and mean CT (mean cortical thickness of all areas) or total surface area (sum of all surface areas). Self-correlations (represented by the main matrix of measurements of the diagonal network) were calculated on the lower triangular part of the matrix. The partial correlation, e _ij (i.e. edge weight) can be calculated according to the following general equations:

Where x _{i ; j} means a matrix of variables, and x _c means any subset of the condition variables. To achieve this general form of partial correlation, the process starts with i, j, c =1, 2, 3:

Therefore, for any subset with conditional variables:

.

.

In some examples, to ensure that the network retained only statistically significant correlations, the calculated correlation coefficients were adjusted for several tests using the Hypothesis Mean False Reject Rate (FDR) procedure outlined in Storey, 2002. The FDR procedure examines each calculated p-value ( from the pairwise correlation calculation) relative to the adjusted significance level, in this example α=0.05, and accepts only p-values less than the adjusted significance level as truly significant. Those pairwise correlations that fail the FDR test can be set to zero; otherwise, all non-zero correlations, both positive and negative, were retained (see Figure 14, described in more detail below).

Thus, it is possible to construct a 68 × 68 correlation matrix for CT or SA in each clinical group, which is a network of structural correlations for either surface area or cortical thickness. The matrix element quantifies the strength of correlation between cortical regions for either cortical thickness or surface area, and does not by itself represent a valid physical connection. In the context of network analysis of structural correlations in neurodegenerative disorders, such correlations are thought to imply either a co-atrophy relationship (if positive) or an inverse atrophy/hypertrophy relationship (if negative) between brain regions.

With respect to the networks and/or structural correlation matrices generated using the above methods, it is useful to use the following metrics to compare the properties of structural networks for the three clinical groups: rib strength, knot grade, intra-module knot grade z-score, and participation index. Edge strength and node degree represent two main network attributes; they respectively quantify the strength of the correlation between nodes and the number of pairwise correlations for each node. To assess whether cortical lobes represent modules, two network metrics are used that measure modularity in network interactions, namely the intra-module degree z-score and participation index. All scores (except node degree) were computed on weighted graphs and evaluated as averages of the four lobes (described below). The scores were computed on either binary or weighted graphs (as discussed below). From purely theoretical studies, it is known that the computed topological properties of the network depend on the choice of the threshold value (van Wijk et al. 2010). In this paper, a fixed threshold is chosen for each group correlation matrix.

Node Degree

The degree of a node, k _i , represents the number of significant correlations for each node in the network. In general, the node degree is computed from binarized correlation matrices, where each significant correlation in the matrix is replaced by either 1 if it is significant or 0 if it is not. Examples of binarized matrices are shown in Figure 16. Binarized matrices may also be referred to as adjacency matrices. In Figure 16, the top three plots correspond to bark thickness and the bottom three plots correspond to surface area. Significant positive correlations are shown in white, while significant inverse correlations are shown in black.

The degree of node i , i.e. the number of meaningful links connected to a node can be calculated as:

Equation 1

where N is the number of nodes and a _ij represents the links between nodes i and j and has a value of 1 if there is a direct link between nodes and 0 if not.

Modularity index

The node participation index and the intra-module degree z-score evaluate the role of the node according to the modules. Network modules (also known as community structures) are densely connected subgraphs of a network, i.e. subsets of nodes within which network connections are denser, and connections between which are more sparse. It is useful to explore the modular organization of the frontal, temporal, parietal, and occipital regions of a network of cortical thickness or surface area defined as modules. Since these cortical surface area lobes are not necessarily modular in themselves, it may be necessary to first check whether the lobes are inherently modular. In one example, this can be done by calculating the modularity index ( Q ) of the networks according to each share. The modularity index quantifies the observed fraction of power values within a module compared to what would be expected if the connections were randomly distributed over the network. Since the constructed crustal thickness and surface area networks contain both positive and negative edge strengths, it is possible to use an asymmetric generalization of the modularity quality function. For example, as presented in Rubinov and Sporns (2011):

Equation 2

равна i, j-му элементу матрицы корреляции, т.е. силе попарной корреляции между корковыми областями, ω _ij , если ω _ij >0, и равна нулю, если нет. Также,

равна -ω _ij , если ω _ij >0, и равна нулю, если нет. Член

означает ожидаемую плотность положительных или отрицательных весов соединений с учетом модели случайных нулей с сохраненной силой, где

и

. Дельта Кронекера δ _{Mi , Mj} равна единице, если i, j-е узлы находятся в одном модуле, и равна нулю, если нет. Эффективность заданного разделения сетей на модули тестировали с применением функции обнаружения сообщества, известной как таковая в данной области техники, при использовании вектора принадлежности узла к конкретному узлу в качестве исходного вектора принадлежности к сообществу.Where

is equal toi, jth element of the correlation matrix, i.e. strength of pairwise correlation between cortical areas,ω _ij , Ifω _ij >0, and zero if not. Also,

equal to -ω _ij , Ifω _ij >0, and zero if not. Member

means the expected density of positive or negative connection weights given the random zeros model with conservation of strength, where

And

. Kronecker deltaδ _{Mi , Mj} equals one ifi, jThe th nodes are in the same module, and is zero if not. The effectiveness of the given division of networks into modules was tested using the community detection function, known as such in the art, using the node-specific membership vector as the initial community membership vector.

It has been found that the lobar organization of the cortical surface into frontal, parietal, temporal, and occipital regions is in fact modular (see Appendix A). Accordingly, it is then possible to calculate the contribution of individual nodes to stake modules as a node participation index and an intra-lobe z-score, which is referred to as an inter-lobe node participation index and an intra-lobe node z-score.

Interlobar Node Participation Index

In general, the participation index p evaluates inter-module connectivity. It can be thought of as the ratio of nodes' intra-partial edges to all other share modules in the network, where node p _i tends to 0 if the node has links exclusively within its module, and tends to 1 if the node communicates exclusively outside its module. The weighted participation of the network is calculated by:

Equation 3

является взвешенным количеством связей i-го узла со всеми другими узлами в модуле m - межмодульной степенью, и

является общей степенью i-го узла. В этом документе, термин междолевое участие применяют для этой меры сети. where M is the set of modules and

is the weighted number of connections of the i -th node with all other nodes in the module m - inter-module degree, and

is the common degree of the i -th node. In this document, the term inter-participation is used for this network measure.

z-score intralobar node degree

Complementing the inter-lobar participation index is the normalized intra-lobar degree, z _j , which assesses intra-lobar connectivity with a z-score, i.e. using the normalized divergence of the interlobar node degree with the corresponding average degree distribution. Therefore, the node's intra-lobe z-score, z _j , is large for a node with a large number of intra-module connections compared to the average inter-module connectivity. For networks where correlation strengths are preserved, the intra-module z-score of the degree of a node is computed as:

Equation 4

такой, как описан выше,

является средним значением внутримодульного m _j распределения степеней, и

является стандартным отклонением внутримодульного m _j распределения степеней.Where

such as described above,

is the mean of the intramodule m _j power distribution, and

is the standard deviation of the intramodule m _j power distribution.

Node Role in Modular Networking

The role of a node with a modular organization of shares depends on its position in the space of the parameter zp _i . There are four possible roles a node can play in a network, which are assigned based on above-average nodal property scores. It is useful to consider two of these roles, the so-called connectors or global network cores (with high inter-participation and high intra-participatory z-score) and the so-called peripheral cores (which have high intra-partial z-score and low inter-participatory participation). Thresholds for high and low z _i and p _i were set above 1.5 and 0.05, respectively.

Statistical analysis

Statistical differences in demographic and cognitive scores among subjects are assessed using either one-way analysis of variance or a two-sample t-test. The data are checked for normal distribution using the one-sample Kolmogorov-Smirnov test. A chi-square test is used to test for differences in the distribution of males and females between groups. Statistical differences in global network correlation strength across diagnostic groups are tested using one-way ANOVA for an unbalanced sample size (to account for an odd number of significant network correlations). Node degree, intralobar z-score, z _i , and interlobar participation index, p _i , are compared between groups using the Kruskal-Wallis test, a non-parametric one-way ANOVA. The results are shown as significant at p <0.05.

results

Table 1 below shows demographic, cognitive, and mean CT and SA for each group according to clinical diagnosis. The 3 groups differed significantly in age, with AD patients being older than HE and bvFTD ( p < 10 ^-4 on all tests). There were also significant differences in MMSE cognitive scores, AD patients had the most impairment, and bvFTD had greater impairment compared to HE subjects ( p <10 ^-4 on all tests). Mean CT and total SA differed between groups.

Table 1

Gender M/F Age (years) average (co) MMSE average (co) CT (mm) average (co) Total SA (×10 ⁵ mm ² ) average (co) HE 106/95 64(1) ^b 29(1) ^{a, b} 2.43(0.1) ^{a, b} 106.9(1.4) ^{a, b} bvFTD 136/77 64(8) ^s 24(4) ^s 2.20(0.15) 102.8(1.5) AD 105/108 71(10) 21(4) ^s 2.18(0.15) 102.5(1.6)

Abbreviations: HE - healthy elderly, bvFTD - behavioral frontotemporal dementia, AD - Alzheimer's disease, M - men, F - women, MMSE - Mini Mental Status Assessment Scale, CT - cortical thickness, SA - surface area. Significant differences between groups: a - HE/bvFTD, b - HE/AD, c -bvFTD/AD ( p <0.05).

Differences between HE and both patient groups were significant for both mean CT ( p <10 ^-4 , both tests) and total SA ( p <0.003, both tests), but the bvFTD and AD groups did not differ from each other. Mean CT and total SA values averaged by brain lobes are shown in Table A.3 in Appendix A. Thus, while AD and bvFTD differ in the proportion of pathology distribution, age, and severity of cognitive impairment, neither the overall degree of cortical thinning, nor is the change in mean surface area a means of distinguishing between the two states.

Structural Correlation Network Share Properties

Since the definition of correlation-based networking depends on the choice of a threshold value, it is useful to ensure that the networks defined here were not random in their global topology by calculating the value of density/sparseness ( κ ). It is believed that brain networks exhibit a non-random (“close world”) topology if κ > 0.1, which was the case for all the networks considered here. It is also useful to ensure that inverse correlations were not missed after thresholding (see Figure 14). Therefore, all the networks of positive and inverse correlation CT and SA considered here are not random. See also Table A.3 for global κ values for CT and SA in the three groups.

Using the modularity index, a study was conducted to determine if cortical lobes, defined as usual, correspond to network modules in the CT network. It was found that only two pairs of homologues (posterior cingulate and precentral cortex) in the CT network and two pairs of homologues in the SA network (posterior cingulate cortex and paracentral and right marginal regions (banks) of the superior temporal sulcus were incorrectly assigned in the modularity index algorithm. In Table A.2 (Appendix A) provides details of the inputs and outputs of the algorithm In practice, a Q value greater than 0.3 is considered to be a good indicator of the presence of significant modules in the network. recalculated for 100 CT matrices generated from surrogate databases Each of the 100 CT and SA surrogate matrices was generated by randomly selecting 213 subjects from three study cohorts and calculating Q values on the correlation matrices generated for CT and SA Q values are shown 15, which is a plot of the distribution of the modularity index Q in regional CT networks created on 100 surrogate datasets. The center line means the mean, and the top and bottom lines mean 1.5 standard deviations from the mean (Q=0.36±0.02), i.e. an arbitrary value, where the Q values are similar to the values of an arbitrary graph. For the study groups, the following values were obtained: Q _HE =0.49, Q _bvFTD = 0.49 and Q _AD = 0.45 for "positive", and Q _HE = 0.39, Q _bvFTD = 0.28 and Q _AD =0.29 for "negative" subnets, which means non-arbitrary modular topological organization in the CT/SA correlation network of the frontal, temporal parietal, and occipital regions.

Thus, it can be concluded that the cortical lobes described as usual correspond to non-random modules in the CT network.

Average correlation strength of CT and SA networks

Figure 5 shows the rib strength values of each cortical thickness correlation network averaged over the brain lobes and compared between the HE, bvFTD and AD groups. Data are shown for networks of positive (upper plot) and inverse (lower plot) correlation networks. Asterisks denote significant differences between the three groups (* p <0.05; ** p <0.01). As seen in Figure 5, the mean strength of correlation for CT showed significant differences between HE, bvFTD, and AD subjects in the frontal, temporal, parietal, and occipital lobes ( p <10 ^-4 for all tests). The mean strength of the correlation was higher in patients with bvFTD and AD than in patients with HE in the frontal, temporal, parietal, and occipital lobes (for all pairwise comparisons, p ≤ 0.003). The mean correlation strength was higher with bvFTD than with AD in the frontal ( p <10-4) and temporal ( p =0.005) lobes.

The mean strength of inverse correlation networks in the CT network also differed in the frontal and temporal lobes, see bottom plot in Figure 5. Again, both bvFTD and AD groups showed higher mean correlation strength than the HE group in the frontal and temporal lobes ( p ≤ 0.003 on all tests), and the bvFTD group had a higher mean inverse correlation strength than the AD group in the frontal lobe ( p = 0.003).

Figure 6 shows that the rib strengths of each surface are a correlation network averaged over the frontal lobe of the brain and compared between the HE, bvFTD and AD groups. Data are shown for networks of positive (right graph) and inverse (left graph) correlations. Asterisks indicate significant differences between the three groups (* p <0.05; ** p <0.01). The graphs show significant differences in the average strength of correlation across the SA network. The diagnostic groups differed only in the frontal lobe, where the AD group had a lower mean correlation strength than the HE group ( p = 0.03). Similarly, SA network inverse correlations were significantly different in the frontal lobe, with lower mean correlation strength in the bvFTD and AD groups compared to the HE group ( p ≤ 0.02 in both tests). This is due to the greater number of correlations with the broader frequency distribution of strength found in disease compared to the sparser networks having a narrower frequency distribution in healthy older adults (see Figure 14).

Measurements of nodes in a CT network

Node Degree

Node grade, which quantifies the average number of significant positive correlations per node, is shown averaged over the frontal, temporal, parietal, and occipital lobes for the CT network in Figure 7. Node grade is compared across the HE, bvFTD, and AD groups. Data are shown for networks of positive (upper plot) and inverse (lower plot) correlation networks. Asterisks indicate significant differences between the three groups (* p <0.05; ** p <0.01).

There were significant differences between groups in the frontal, temporal, parietal and occipital lobes ( p ≤10 ^-4 in all tests). Subjects with both bvFTD and AD had a higher degree of nodule in the frontal and temporal lobes ( p < 0.006 for all tests) compared to HE subjects. The bvFTD group had a markedly higher degree of knot in the parietal and occipital lobes than the AD group ( p ≤ 0.02 for all tests). A similar pattern was found for the number of inverse correlations in the CT network in the frontal, temporal, parietal, and occipital lobes ( p ≤ 0.02 in all tests). These differences were due to more significant inverse correlations in bvFTD and AD than in the HE group across all four lobes ( p <0.01 on all tests). None of the differences between the bvFTD and AD groups were significant.

Interlobar Node Participation Index

Group differences were found in the interlobar index of node involvement for CT. The index measures the degree of significant positive correlation with nodes in different proportions. They were significant for the lobes located in the temporal, parietal, and occipital lobes ( p ≤ 0.03 for all tests). The differences reflect higher index values compared to the HE group in the parietal ( p <0.003 in both groups), temporal ( p =0.01 in AD), and occipital ( p =0.002 in bvFTD) lobes. This is shown in Figure 8, where the interlobar index of node participation in the cortical thickness correlation network, averaged across the lobes, is compared in the HE, bvFTD, and AD groups. Data on the graph is shown only for positive correlations. Asterisks indicate significant differences between the three groups (* p <0.05; ** p <0.01). None of the inter-lobe participation index comparisons differed significantly in any of the lobes for inverse correlation in the CT network.

Measurements of nodes in the SA network

Node Degree

SA network node degree values are shown for the frontal, temporal, parietal, and occipital lobes in Figure 9. In the figure, surface area correlation network node degree is averaged across the brain lobes and compared across the HE, bvFTD, and AD groups. Data are shown for networks of positive (upper plot) and inverse (lower plot) correlation networks. Asterisks indicate significant differences between the three groups (* p <0.05; ** p <0.01).

Positive correlations differed between diagnostic groups in the frontal, temporal, parietal, and occipital lobes ( p ≤ 0.03). As with the CT network, the bvFTD and AD groups had a higher degree of SA node than the HE group in the frontal, temporal, and parietal lobes ( p < 10 ^-4 on all tests). For the occipital lobe, the only difference that was significant was between the AD and HE groups ( p =0.04). In contrast to the CT network, the degree of node in the parietal lobe was also significantly higher in AD than in bvFTD ( p = 0.004).

The SA inverse correlation network also showed significant group differences in the frontal, temporal, parietal, and occipital lobes ( p ≤ 0.001 in all tests). Again, both the bvFTD group and the AD group had higher nodal grades than the HE group in all four lobes ( p < 0.001 for all tests). In contrast to the CT inverse correlation network, the AD group had a higher node degree than the bvFTD group in the frontal ( p =0.02) and parietal ( p =0.01) lobes.

Interlobar Node Participation Index

Figure 10 shows the group differences in the inter-participation node participation index for the network organization SA. The figure shows interlobar nodal involvement indices averaged across brain lobes and compared across the HE, bvFTD and AD groups. Data are shown for networks of positive (upper plot) and inverse (lower plot) correlation networks. Asterisks indicate significant differences between the three groups (* p <0.05; ** p <0.01).

Both the bvFTD and AD groups had higher index values than the HE group for the SA positive correlation network in all four lobes ( p < 10 ^-4 ). In contrast to the CT correlation network, the SA inverse correlation network also showed significant differences in the frontal and parietal lobes ( p ≤ 0.04 for both patient groups) and in the temporal lobe for the AD group ( p < 0.001) relative to the HE group.

Kernels of Structural Correlation Networks

CT network cores

There are four possible combinations of mean values for the interlobar participation index (p high/low) and intralobar z-score (high/low). Here, only the case of high interlobar index and high intralobe z-score is considered in order to focus on nodes with high kernel-like characteristics. Tables A.4-A.6 (see Annex A) provide data for the cores of the global and edge networks. The remaining two combinations were studied but were not informative. The number and distribution of network cores within high p and z values in the CT positive correlation network differed between study groups. In HE subjects, the nuclei were distributed throughout the cortex; each lobe had at least one nucleus, with four nuclei in the frontal lobe. The reorganization of the nuclear topology occurred differently in the two groups of diseases. This is shown in Figure 11, top panel, which is a visualization of the cores of the cortical thickness network in the brain space. In bvFTD, the number of nuclei in the frontal lobe increased from 4 to 9, decreased from 2 to 1 in the occipital lobe, and completely disappeared in the parietal and superior lobes. In contrast, the cores were distributed across all four AD lobes almost equally. The number of nuclei in the frontal lobe decreased (2 vs. 4), while the number increased compared to the HE group in the temporal and occipital lobes (1 vs. 3 and 2 vs. 3, respectively). A complete list of cores that have core-like properties in the CT network, along with stake positions, is presented in Table A.4 (see Appendix A).

Nodes with core-like properties in the CT inverse correlation matrix were exclusively present in the frontal and temporal lobes in all three groups, and their topological distribution differed between groups. See bottom panel of Figure 11 and Table A.4 in Annex A.

SA network cores

The cores in the SA positive correlation network are shown in Figure 12, top panel. Table A.5 (see Appendix A) provides a list of nodes and stake positions classified according to inter-lobe participation index and intra-lobe z-score. A visual comparison of the nuclear topology between groups shows that there were more nodes with nucleus-like properties in the left hemisphere in all diagnostic groups. However, HE subjects had only one SA (left cerebral insula) nucleus, whereas both disease groups had more nuclei in each lobe. There were twice as many SA nuclei in the AD group as compared to bvFTD (14 versus 7). Unexpectedly, the bvFTD group had more SA nuclei in the temporal than in the frontal lobe (4 vs. 1), while the AD subject had more frontal than temporal nuclei (6 vs. 4). Subjects with AD had 3 nuclei in the parietal lobe compared to 1 in subjects with bvFTD.

The nuclei in the SA inverse correlation network are present in either the frontal or temporal lobes only in all three groups. However, the HE group had one nucleus in the parietal (precuneus) and bvFTD had two (infero-parietal and paracentral) (see Table A.5 in Appendix A). Interestingly, most SA inverse correlation nuclei in AD were found in the frontal lobe.

Topology of combination of bark thickness - bark surface area

The strength of the combination between CT and SA nodes was calculated by elementwise multiplication of the respective correlation matrices of CT and SA. Figure 13 shows the strength of the CT/SA combination visualized in brain space. It can be seen that in HE subjects, the interhemispheric homologue pairs show a conjugate correlation of CT/SA. On the contrary, the combination of CT/SA in the AD and bvFTD groups is very similar to each other and differs from the HE group. Both bvFTD and AD groups showed greater association between non-homologous nodes in the same and contralateral hemispheres. Interlobar correlations were also strikingly different between the bvFTD and AD groups. In the bvFTD group, most of the interlobar CT/SA correlations were due to frontotemporal interactions. In AD, most of the interlobar CT/SA interaction was due to fronto-parietal interactions. A list of CT/SA combination topology cores is given in Table A.6 of Annex A.

Discussion

Baseline structural correlation networks in subjects clinically diagnosed with bvFTD or AD in three large global clinical trials were examined and compared with healthy older subjects in a well-characterized age cohort. For each group, networks were constructed based on partial correlations between 68×68 pairs of cortical surface regions (nodes) in terms of their thickness and surface area. The approach adopted allowed a systematic analysis of both positive and inverse network correlations in three clinical contexts. The methods and data discussed herein represent the first systematic comparative analysis of cortical thickness and surface area in a large population of subjects. Since the number of patients in the three groups should have been comparable, the overall study size is determined by the number of available subjects with bvFTD. Because it is a rare disease, it was necessary that the bvFTD component of the study be global, with patients coming from 70 study sites in 13 countries. For the 213 patients included in the study, this represents the largest array of MRI data in subjects with bvFTD currently available. To match this, 213 patients were randomly selected from a much larger cohort of 1131 AD patients coming from 116 centers in 12 countries for study TRx-237-005 and 128 centers in 16 countries for study TRx-237-015 (available, e.g. from the US National Library of Medicine). The 202 normal elderly come from a well-characterized age cohort that was studied longitudinally. Thus, the results presented are reliable and can be considered representative of an international population meeting accepted diagnostic criteria.

Modularity of networks by shares

Structural correlations in the frontal, temporal, parietal, and occipital cortical surfaces have been shown to be inherently modular for both networks, both cortical thickness and surface area. That is, the results confirm that the standard lobar cortex shares common attributes of network modularity, so that they differ from what would be expected in a comparable random network. Highly clustered network modules impart the properties of a so-called "small world" network, and are thought to strike an optimal balance between local specialization and global integration. The results in healthy older adults are comparable to previous studies in a smaller and younger healthy group, revealing an underlying modular architecture in the Regional Thickness Correlation Network. The results also indicate that the inherent lobe-like modularity is preserved in both bvFTD and AD, indicating that the overall lobar network architecture is preserved in the presence of neurodegenerative changes. As discussed below, this is in contrast to the core-like networking that varies with disease.

Similarities and differences between AD and bvFTD relative to healthy elderly subjects

The morphological correlation networks for both patient groups (bvFTD and AD) were found to be very different from the corresponding network for healthy older adults. Both groups showed a striking increase in the overall strength of correlation in thickness and surface area networks compared to healthy older adults. The action was more pronounced in the cortical thickness network in all lobes for both positive and inverse correlations. This contrasts with the significantly lower than normal correlation strength for frontal surface area in AD and the similarly directional difference in bvFTD. This may be due to the greater number of correlations with a wide frequency distribution in disease compared to the sparser networks with a narrower frequency distribution in healthy older adults. In addition to an increase in overall correlation strength, the number of intralobar positive and inverse correlations as measured by node degree was higher in all lobes in both dementia groups than in control healthy older adults. The number of interlobar positive thickness correlations, as measured by the interlobar participation index, was also higher in all lobes. The number of intralobar and interlobar positive surface area correlations was also greater in both bvFTD and AD than in healthy older adults in the frontal, temporal, and parietal lobes. Both disease groups also differed from healthy older adults in terms of correlations in the combination between cortical thickness and surface area. Thus, both diseases are characterized by a general increase in the strength and degree of structural correlation, occurring both locally within the lobes and globally between the lobes.

The similarity between the two conditions in terms of a marked increase in overall strength and degree of structural correlation may appear to call into question the clinical differences between bvFTD and AD on which the subject classification in the study was based. Indeed, there were no differences between the two states in terms of overall crustal thickness and surface area. However, there are some important network differences between the two states. In the cortical thickness network, the overall positive correlation strength was greater in bvFTD than in frontal and temporal AD, and the inverse correlation strength was also greater in bvFTD than in frontal AD. The number of significant positive intralobar correlations was higher with bvFTD than with AD in the parietal and occipital lobes. Conversely, the number of positive and inverse intralobar correlations was greater in AD than in bvFTD in the frontal and parietal lobes. Most of the inverse correlations in cortical thickness and surface area are associated with interhemispheric non-homologous frontotemporal lobes in bvFTD and frontoparietal lobes in AD.

The core-like organization of the correlation networks also differed significantly in these two states. However, network connector cores are considered to provide network integration, while edge cores provide network segregation. It has been suggested that the nuclei provide adaptability to stroke in neurodegenerative disorders. Alternatively, it has been suggested that the cores represent the loci of a particular vulnerability. Therefore, it would be interesting to study how nuclei change in the context of a neurodegenerative disease. bvFTD is characterized by an increase in the number of nuclei of cortical thickness in the frontal lobe and a decrease or elimination of nuclei in the temporal, parietal, and occipital lobes. Conversely, AD is characterized by nuclei distributed in all lobes, a decrease in the number of nuclei in the frontal cortex, and an increase in the number of nuclei in the temporal and occipital lobes compared to bvFTD. In a network of positive surface area correlation, AD subjects had twice as many nuclei as in bvFTD, and the topology of these nuclei differed. Thus, in general, AD is characterized by a much more distributed structure of nuclei in degenerative networks of both thickness and surface area than bvFTD. Conversely, the nucleus-like organization is much more localized in bvFTD. bvFTD is said to be a clinical syndrome with focal but heterogeneous atrophy centered around the nuclei. The identification of the insular region as one of the reverse network nuclei (both in the bvFTD and AD groups for the CT network) is consistent with a recent and unexpected finding from diffusion MRI that there is an increase in insular nucleus-like fiber connectivity in bvFTD. The nuclei of the healthy elderly group, on the other hand, were closely related within and between lobes in a homologous manner and were not otherwise related to each other. Differences in core-like organization between AD and bvFTD indicate differences in the hierarchy of node vulnerability and in the organization of network adaptations that compensate differently in the two states. Thus, in contrast to the lobar modularity that is preserved in neurodegenerative diseases, a permanent nucleus-like organization is not preserved, implying that pre-existing nuclei are not an inherent structural property of cortical network organization.

Although AD is also characterized by changes in cortical thickness, these are generally less pronounced than in bvFTD, while changes in surface area are more prominent in AD, suggesting coordinated changes in the number of adjacent affected columns. These differences are consistent with the bvFTD pathology affecting interneurons and astrocytes, which have more localized connections. The predominance of surface area correlations in AD is consistent with the pathology primarily affecting the conductive cortico-cortical projection systems mediated by basal cells. bvFTD differs from AD in many important respects: bvFTD has no cholinergic deficit, no therapeutic effect from treatment with acetylcholinesterase inhibitors or memantine, bvFTD is characterized by severe astrocytic pathology, neurons affected in the neocortex are predominantly spiny interneurons in layers II and VI (pyramidal cells in layers III and V are predominantly affected in AD) and the dentate gyrus of the hippocampus (neurons affected in AD are in CA 1-4 rather than in the dentate gyrus), and bvFTD is characterized by increased levels of glutamate in the neocortex, while AD is not. However, none of these conditions provide a simple explanation for the various patterns of distribution of correlated structural changes described herein.

The global nature and significance of network changes in dementia

The overall picture that emerges from the two disease groups studied is that the network architecture changes in a coordinated manner throughout the brain with respect to both positive and inverse correlations. This is surprising given that the neurodegenerative processes in these two conditions are generally considered to be anatomically limited to the frontal and temporal lobes in the case of bvFTD and to the temporal and parietal lobes in AD. Rather, network analysis suggests that in both states there are changes in networks of cortical thickness and surface area that affect all lobes globally, but there are differences in the anatomical topology of the changes. The aggregation pathology of both tau and TDP-43 is known to propagate in a prion-like manner, whereby pathology in an affected population of neurons can initiate pathology in an associated but previously unaffected population of neurons. Therefore, positive correlations may partly reflect the spread of pathology in existing normal networks, whereby existing functional networks are affected or kept together. Alternatively, such correlations may express functional dependencies such that a loss of function in one member of a partnership results in a parallel loss of function in a partner, typically functionally synchronized with the affected node. This interpretation is consistent with previous work on cortical thickness correlations in healthy adults, where positive correlations were found to converge with diffusion-based axon connections.

The work discussed here highlights for the first time the importance of inverse correlation networks. It should be noted that since the inverse correlations observed in both neurodegenerative disorders primarily reflect interlobar non-homologous associations, they would not have been detected using a share-only analysis approach. In particular, the emergence of these non-homologous inverse interlobar correlations and their increased strength represent the clearest overall difference between neurodegenerative disease and normal aging. Conversely, the normal aging brain is characterized by significantly weaker homologous positive correlations. An attractive hypothesis is that as certain nodes become functionally impaired, other unaffected nodes compensate by emphasizing non-homologous associations in the disease. This would mean that the observed basic reorganization of the structural network could be partly adaptive. Structural plasticity has been demonstrated in other contexts, and functional compensation is known to occur in focal disease.

The work discussed here represents the first comparative study of structural correlation network anomalies in bvFTD and AD compared with healthy aging. These correlations arise from both positive and inversely related changes in cortical thickness and surface area in two disease states that are very different from those seen in normal older adults. The changes observed in the disease are global in nature and are not limited to the frontotemporal and temporoparietal lobes, respectively, in bvFTD and AD. Rather, they appear to represent structural adaptations to neurodegeneration that differ between the two conditions. In addition, all correlation networks showed a very different core-like organization that differs from both normal and the two forms of dementia. In contrast to the lobar organization of networks, which remains constant during disease, the nucleus-like organization changes depending on the underlying pathology. This means that nucleus-like organization is not a fixed feature of the brain, and attempts to explain disease in terms of nuclei may be inadequate. The documented differences between AD and bvFTD confirm that the clinical differences in the two populations with dementia correspond to systematic differences in the underlying network structure of the cerebral cortex. The topological differences in thickness and surface area in the nucleus-like organization, and the underlying network of positive and inverse correlations, could form the basis for the development of analytical tools to aid in the differential diagnosis of these two conditions, which are difficult to distinguish on purely clinical criteria.

Application of correlation matrices to determine the response of a group of patients to neuropharmacological intervention

The methods discussed above are used to determine the response of a group of patients to a neuropharmacological intervention.

Figures 17A-17D depict correlation matrices for two patient populations: those treated with symptomatic AD drugs (cholinesterase inhibitor and/or memantine; ach1 in the legends) and those not being treated (ach0 in the legends). Subjects are ranked on a clinical rating scale for dementia (CDR) of 0.5, 1, or 2. Figure 17A presents a cortical thickness correlation matrix at baseline (i.e., week 0) for 96 subjects diagnosed with AD who received no symptomatic treatment . Conversely, Figure 17B presents a correlation matrix of cortical thickness at baseline for 445 subjects diagnosed with AD who are receiving symptomatic treatment. Figure 17C is a surface area correlation matrix at baseline for 96 subjects diagnosed with AD who were not treated symptomatically, and Figure 17D is a surface area correlation matrix at baseline for 445 AD patients who were treated symptomatically.

As seen in Figures 17A-17D, symptomatic treatment of AD causes a significant increase in interlobar non-homologous inverse correlation networks (blue in Figures 17B and 17D) compared to untreated patients (Figures 17A and 17C). This is especially noticeable for surface area networks.

These associations are inverse correlations, where a decrease in the volume or surface area of the affected area in a particular node (usually located in the posterior regions of the brain) is statistically significantly correlated with the associated node, where there is a corresponding increase in volume or surface area. As discussed above, the presence of these non-homologous inverse correlations is indicative of a neurodegenerative disease and most likely represents frontal compensation for posterior dysfunction resulting from the pathology. Symptomatic treatment of AD causes an increase in these non-homologous compensatory connections.

Figures 18A-18D are plots of non-homologous interlobar node degrees (as discussed above) for cortical thickness-positive correlations, cortical thickness-inverse correlations, surface area-positive correlations, and surface area-inverse correlations, respectively. As can be seen from these graphs, the number of significant non-homologous interlobar compensatory inverse correlations increases significantly with symptomatic treatment of AD.

Figures 19A and 19B show cortical thickness correlation matrices based on structural neurological data that are temporally separated. Figure 19A is a cortical thickness correlation matrix at week 0 (ie, at baseline) for a group of 445 patients diagnosed with AD who are receiving symptomatic treatment for AD. Figure 19B is a cortical thickness correlation matrix at 65 weeks for the same cohort of 445 patients diagnosed with AD. During the interim period, the group was also treated with leucomethylthioninium mesylate (LMTM; USAN name: hydromethylthionine mesylate), a tau protein aggregation inhibitor, at a dosage of 8 mg/day (hereinafter 4 mg twice daily). As can be seen, LMTM has minimal impact on structural correlation networks in patients receiving symptomatic AD treatment.

Figures 20A-20D are graphs of non-homologous interlobar nodule grade compared between week 0 and week 65 in the ach1 group (simultaneously receiving symptomatic AD treatment) for, respectively, cortical thickness - positive correlations, cortical thickness - inverse correlations, surface area - positive correlations and surface area - inverse correlations. As can be seen, there is a minimal overall effect of LMTM, as a supplement, on brain network correlation structures at 65 weeks. Importantly, this is an intra-cohort analysis with patients at baseline serving as their own control for changes occurring after 65 weeks of LMTM treatment.

Figures 21A and 21B show cortical thickness correlation matrices based on structural neurological data that are temporally separated. Figure 21A is a cortical thickness correlation matrix at week 0 (ie, at baseline) for a cohort of 96 AD patients treated with LMTM as monotherapy at a dosage of 8 mg/day. Figure 21B is a cortical thickness correlation matrix at 65 weeks for the same cohort of 96 patients diagnosed with AD. Ninety-six patients in this cohort received no symptomatic treatment for AD in combination with LMTM. As can be seen, LMTM as a monotherapy results in a significant reduction in thickness correlations, both intralobar (positive) and interlobar compensatory (inverse) correlations. This is an intra-cohort analysis where patients at baseline serve as their own control for changes occurring after 65 weeks of LMTM treatment.

Figures 22A and 22B are graphs of the interlobar node degree compared between week 0 and week 65 in the ach0 group for cortical thickness - positive correlations and cortical thickness - inverse correlations. The graphs show a very significant effect of LMTM at 8 mg/day as monotherapy on the number of interlobar correlations in the AD group. A significant decrease in the number of positive and inverse non-homologous correlations of cortical thickness is observed after 65 weeks. This is likely due to the normalization of neuronal function in the posterior regions of the brain, whereby LMTM reduces pathology and reduces neuronal dysfunction due to pathology, thereby reducing the need for compensatory input from unaffected or less affected frontal areas of the brain.

Figures 23A and 23B show surface area and thickness correlation matrices based on structural neurological data that are temporally separated. Figure 23A shows the surface area correlation matrix at week 0 (ie, baseline) for a group of 96 patients diagnosed with AD who continued to take LMTM as monotherapy at a dosage of 8 mg/day. Figure 23B shows the surface area correlation matrix at week 65 for the same cohort of 96 patients diagnosed with AD. Ninety-six patients in this cohort were not receiving concomitant symptomatic AD treatment. As can be seen, LMTM as monotherapy results in a significant reduction in surface area correlations, both intralobar (positive) and interlobar compensatory (inverse) correlations. This is an intra-cohort analysis in which patients at baseline serve as their own control for changes occurring after 65 weeks of LMTM treatment.

Figures 24A and 24B are plots of non-homologous interlobar node degree compared between week 0 and week 65 in the ach0 group for surface area positive correlations and surface area negative correlations. Graphs show a significant effect of 8 mg/day LMTM as monotherapy on the number of interlobar correlations in the AD group. Notably, after 65 weeks, there is a significant reduction in the number of positive and inverse/compensatory surface area correlations.

Figures 25A-25D show cortical thickness correlation matrices compared between 96 AD patients (with CDR 0.5, 1, or 2) at baseline and at 65 weeks versus 202 healthy elderly controls. As can be seen, LMTM at a dose of 8 mg/day as monotherapy approximates the cortical thickness networks to normal.

Figures 26A-26D show surface area correlation matrices compared between 96 AD patients (with CDR 0.5, 1, or 2) at baseline and at week 65 compared to a control group of 202 healthy elderly patients. As can be seen, LMTM at a dose of 8 mg/day as monotherapy normalizes surface area networks.

Figures 27A-27D show cortical thickness correlation matrices compared between a group of 54 AD patients with a CDR of 0.5 at baseline alone and at 65 weeks compared to a control group of 202 healthy elderly patients. As can be seen, LMTM at a dose of 8 mg/day as monotherapy reduces the number of inverse/compensatory non-homologous correlations, which become equivalent to the normal control of old age.

Figures 28A-28D show surface area correlation matrices compared between a group of 54 AD patients with a CDR of 0.5 at baseline alone and at week 65 compared to a control group of 202 healthy elderly patients. As can be seen, LMTM at a dose of 8 mg/day as monotherapy reduces the number of inverse/compensatory non-homologous correlations to be equivalent or lower than in a normal control group of elderly people.

In general, the structural correlation network analysis discussed above shows the occurrence of highly anomalous inverse non-homologous interlobar correlations in AD and bvFTD. It is hypothesized that they represent a compensatory contribution from the frontal areas of the brain not affected or less affected by the disease. Symptomatic treatment and LMTM act fundamentally differently in AD in terms of a network of structural correlations. Symptomatic treatment causes a significant increase in compensatory networks. LMTM as a monotherapy reduces the need for these compensatory networks by reducing primary pathology, thereby allowing affected neurons to function more normally. These results confirm that the abnormal inverse non-homologous correlations observed in neurodegenerative diseases such as AD are adaptive in that they can be reversed or attenuated by a disease-modifying treatment, but not by a symptomatic AD treatment. Effects are observed within a pre/post cohort analysis in which subjects at baseline serve as their own control for changes occurring after treatment with LMTM at 8 mg/day as monotherapy for 65 weeks. These analyzes are much more sensitive to the effects of treatment than are rough analyzes of whole brain volume or lobes of the brain. In addition, as will be discussed below, the results obtained in terms of structural correlation networks are consistent with the functional effects observed with renormalized partial directional coherence electroencephalography analysis methods.

Structural/functional correlation using electroencephalography (EEG)

Renormalizable partial directional coherence (rPDC) network approaches to EEG data allow determination of the direction and strength of electrical activity in the brain, which should be investigated using a network approach. This is discussed, for example, in WO 2017/118733 (whose full content is incorporated herein at this link). Figure 29 shows an example of raw EEG data, and Figure 30 shows an example of an rPDC network derived from the collected EEG data.

The resulting network, as shown in Figure 30, contains a number of nodes indicating approximate locations in the brain (the figure is drawn schematically looking down at the head, with the triangle at the top indicating the nose). The location of the nodes is determined by the placement of electrodes on the surface of the scalp, which are used to obtain EEG data, such as those shown in Figure 29. Directional connections between nodes indicate the flow of electrical activity from one node to another within the brain.

By counting the number of directional connections to and from a given node and/or measuring their relative strength, it can be determined whether the node is an acceptor (and has more and/or stronger connections at the input than at the output) or a donor (and has more and /or stronger connections at the output than at the input). This is shown schematically in Figure 31 where the number/strength of incoming directional connections is subtracted from the number/strength of outgoing directional connections. Thus, in the extreme case, if the difference is negative, then the node acts as a net donor, and if it is positive, the node acts as a network acceptor. More generally, as can be seen in the graph shown in Figure 40, lower values indicate more/stronger outgoing connections and higher values indicate more/stronger incoming connections.

After determining the difference between the incoming and outgoing connection for all nodes, a heat map can then be obtained showing the location and intensity of acceptors and donors in the patient's brain. This may include the step of identifying each node as an acceptor or donor. An example of such a heatmap is shown in Figure 32. In this example, blue areas (arrow A) indicate more outgoing connections and therefore contain more donor nodes, while red/yellow areas (arrow B) indicate more incoming connections and , therefore, contain more acceptor nodes. This type of heat map can be called a "brain print".

Figure 33 illustrates the asymmetry visualization on the heat map from Figure 32, which compares the number of donors and acceptors on both sides. A higher difference of donors and acceptors between the left and right sides of the heatmap is shown in yellow (arrow A), while a smaller difference is shown in black (arrow B).

The methods discussed above are used to analyze data from 329 subjects divided into 167 diagnosed subjects (DS) and 162 paired volunteers (PV) at their initial evaluation (Visit 1):

DS AVERAGE (CO) PV AVERAGE (SD) Age (at 1 visit) 70.23 (9.10) 69.18 (11.21) Floor M: 80, W: 87 M: 75, W: 87 MMSE 1 visit 23.29 (2.60) 28.80 (0.99) ADAS-Cog V1 16.26 (7.21) 6.04 (2.81) CDR 0.62 (0.22) 0.06 (0.16)

MMSE - Mini Mental Status Scale; ADAS-Cog - Alzheimer's Disease Assessment Scale - Cognitive Subscale

As can be seen, the diagnosed subjects have significantly more cognitive impairment on the MMSE and ADAS-Cog psychometric scales, and they also score higher on the overall dementia clinical rating scale (CDR). Otherwise, there are no differences in age or sex distribution.

Figure 34 shows a heat map showing the location of acceptors and donors in the brain of a group of diagnosed subjects at baseline. Arrow A indicates a region of blue regions that contain more/stronger donors and arrow B indicates red regions that contain more/stronger acceptors. Figure 35 is a heat map showing the location of acceptors and donors in the brain of a group of paired volunteers. When comparing the two images, it is clear that AD patients have significantly stronger donors (i.e., more/stronger outgoing connections, shown in blue) in their frontal lobes and significantly stronger acceptors (i.e., more/stronger input connections, shown in red/orange) in the posterior parietal, temporal, and occipital lobes than paired volunteers.

The machine learning classifier was trained on the dataset provided by the 329 subjects described above. In each case, β-range EEG data from 100 seconds of brain activity at rest with eyes closed was used to prepare the rPDC network. Then, using a machine learning classifier, all 329 subjects were classified as either AD or Paired Volunteers (PV) with 95% accuracy. Moreover, a machine learning classifier can be used to estimate the probability that a subject has AD, allowing more than just a binary decision to be made. For example, the subject whose heatmap is shown in Figure 36 has AD. The fact that the patient has AD is revealed at clinical diagnosis. The machine learning classifier estimated with 99% probability that the patient had AD and thus correctly classified this subject. Figure 37 is another example of a heatmap of a patient who has AD as a result of clinical diagnosis. In this case, the machine learning classifier estimated that there is a 63% chance that the patient has AD and (hence) a 37% chance that the patient does not have AD. This information can be used to determine a patient's predisposition to AD if a clinical diagnosis is not made.

Moreover, a specific distribution pattern of abnormal acceptor regions that is indicative of underlying dysfunction can be correlated to specific clinical testing patterns for further more detailed neuropsychological testing and clinical evaluation in the future. For example, the case illustrated in Figure 36 may have a form of dementia other than AD, although classified here as suffering from AD.

Psychometric testing of an apparently healthy cohort showed a downward cognitive trajectory on the Hopkins Auditory Learning Test over 18 months in a subset of subjects. The characteristics of the cohort were as follows. As can be seen, there was no difference in baseline MMSE cognitive scores between those who were at risk and those who were not.

At risk, average (SD) Not at risk, moderate (SD) N 15 88 Age 69.2 (5.82) 67.43 (4.69) Floor 7 women, 8 men 58 women, 30 men MMSE 28.5 (1.74) 28.84 (1.32) years of education 11.73 (1.37) 11.64 (1.23)

A heat map of the group of subjects at risk is shown in Figure 38, while a heat map of the group of subjects not at risk is shown in Figure 39. Since both groups are from an apparently healthy cohort, the differences here are not as clear as for the AD groups. and PV above. Figure 38 shows more/stronger acceptors in the posterior regions of the brain, visible as more intense red/orange in the heatmap. Figure 40 is a box and whisker plot comparing donors and acceptors in frontal and posterior EEG networks at the group level. As can be seen from this figure, the risk group is characterized by increased outward activity from the frontal cortex and increased inward activity in the posterior regions of the brain. EEG recordings were made at baseline, to any measurable decline based on the Hopkins Auditory Learning Test. Therefore, an apparently normal subject who is at risk of deterioration over the next 18 months can be identified already at baseline based on a heat map of his brain activity obtained in a non-invasive way using EEG analysis.

As shown above, there are clear differences in networks between diagnosed subjects and paired volunteers. These differences are very significant at the group level. It should be understood that the first version of the machine learning classifier has a higher level of accuracy than the usual superficial clinical assessment and gives the probability of having AD at the individual subject level, which can be used to make decisions in further clinical management.

Figure 41 shows a comparison of cortical thickness correlation matrices (discussed above) at week 0 for the ach0 AD group compared to a heat map of group-diagnosed subjects. As can be seen, there are a significant number of inverse non-homologous correlations between the frontal lobe and the posterior parietal and occipital regions of the brain. A heat map of a group of diagnosed subjects shows the same phenomenon in terms of brain connectivity as measured by EEG. Both structural and EEG approaches show the same pattern of increased activity from frontal to posterior. Figure 42 shows a comparison of the cortical thickness correlation matrix (as discussed above) at week 0 for a healthy elderly group compared to a group level paired volunteer heatmap. As can be seen, the absence of any inverse non-homologous correlations between the frontal, posterior parietal, and occipital regions of the brain is consistent on the EEG with the absence of increased electrical activity from the anterior to posterior region.

The increase in non-homologous interlobar compensatory inverse correlation networks shown in Figure 41 provides the structural basis for characteristic heatmap changes seen as EEG functional changes.

Figure 43 is a box and whisker plot showing the quantitative differentiation of moderate AD from older adults in the control group. As can be seen, subjects with AD have more outbound activity from the frontal cortex and more inbound activity into the posterior cortex in the β band.

Figure 44 shows three heat maps, from left to right: a group of patients diagnosed with AD who received symptomatic treatment (med), a group of patients diagnosed with AD who did not receive symptomatic treatment (nonMed), and a group of paired volunteers. Figure 45 shows a box-and-whisker plot comparing networks for drug groups, no drugs, and paired volunteers. These data are derived from a preliminary study of 53 diagnosed (DS) patients, of whom 15 received standard drugs and 38 did not receive standard drugs. The characteristics of the two groups are shown in the table below. Although the non-medicated group is significantly younger, there is no difference between the two groups in terms of MMSE-measured cognitive scores or sex distribution.

Med nonMed AGE 75.2 (7.6) 67.5 (9.5) *** FLOOR 8 W: 7 M 23 W: 15 M MMSE 23.84 (1.95) (N=13) 23.37 (2.41)

***p<0.005

There were also no statistically significant differences in the ADAS-Cog or CDR scores.

As seen in Figures 44 and 45, both groups of AD subjects have greater frontal cortex activity in the β range than paired volunteers. In addition, it can be seen that the symptomatic treatment increases activity emanating from the frontal lobe compared to the non-medicated group. This is shown in the box and whisker plot of Figure 45. The drug group has significantly more outgoing electrical activity from the frontal cortex. In the posterior regions of the brain, symptomatic treatment reduces the need for maintenance input electrical activity.

The frontal lobes show the same EEG phenomenon as the results of structural analysis of correlation networks in Figures 17A-D and 18A-D. Figure 44 shows that differences that can be detected at the group level using structural analysis of MRI can also be detected using EEG. It should be noted that although structural analysis of network differences between patients receiving and not receiving symptomatic treatment suggests an increase in non-homologous interlobar connectivity directed to the posterior regions of the brain, EEG analysis shows less input activity directed to the posterior regions. It is currently assumed that symptomatic treatment increases the incoming activity in the posterior regions of the brain in other frequency bands.

The systems and methods of the above embodiments may be implemented in a computer system (eg, computer hardware or computer software) in addition to the structural components and user interactions described.

The term "computer system" includes computer hardware, software, and data storage devices for implementing the system or executing the method according to the above-described embodiments. For example, a computer system may include a central processing unit (CPU), input means, output means, and data storage. Preferably, the computer system has a monitor for displaying a visual image. The data store may contain RAM, disk drives, or other machine-readable media. A computer system may include a plurality of computing devices connected to a network and capable of communicating with each other over the network.

The methods of the above embodiments may be provided as computer programs, or as computer program products or computer-readable media containing a computer program that, when executed on a computer, is configured to perform the method(s) described above.

The term "computer-readable media" includes, without limitation, any non-volatile media or media that can be read and accessed directly by a computer or computer system. Media may include, but is not limited to, magnetic storage media such as floppy disks, hard disk media, and magnetic tape; optical storage media such as optical discs or CD-OM; electrical storage media such as memory including RAM, ROM and flash memory; and hybrids and combinations of the above, such as magnetic/optical storage media.

Although the invention has been described in connection with the exemplary embodiments described above, many equivalent modifications and variations will be apparent to those skilled in the art upon receipt of this description. Accordingly, the exemplary embodiments of the invention set forth above are considered illustrative and not restrictive. In the described embodiments, various changes can be made without departing from the essence and scope of the invention.

In particular, while the methods of the above embodiments have been described as being implemented in the systems of the described embodiments, the methods and systems of the present invention need not be implemented in combination with each other, but may be implemented on alternative systems or using alternative methods, respectively.

APPENDIX A

Table A.1 - Areas of the cortical surface of the frontal, temporal, parietal or occipital lobes according to the Desikan-Killiany Atlas (DKA). The cortical areas (nodes) of each structural correlation matrix are ordered according to the list below throughout the document:

Region designation Share Caudal anterior cingulate Frontal Caudal middle frontal Frontal frontal pole Frontal Island Frontal Isthmus of the cingulate gyrus Frontal Lateral orbito-frontal Frontal Medial orbitofrontal Frontal tire part Frontal Orbital part Frontal triangular part Frontal Precentral Frontal Rostral anterior cingulate Frontal Rostral-median frontal Frontal Upper frontal Frontal banksta temporal Entorial temporal Fusiform temporal Temporal inferior temporal Temporal average temporal parahippocampal temporal Temporal superior temporal temporal pole temporal Temporal transverse temporal Inferior parietal Parietal Paracentral Parietal Postcentral Parietal Back waist Parietal Fore cunee Parietal Superior parietal Parietal Supramarginal Parietal Wedge Occipital Lateral occipital Occipital lingual Occipital Pericalkarin Occipital

Table A.2 - Node assignment to frontal, temporal, parietal, or occipital lobes by algorithm and DKA cortical fragmentation. * nodes are erroneously assigned to a stake for a network of bark thickness, and ⁺ nodes are those that are incorrectly assigned to the surface area network

CT network SA network Region Algorithm tagging DKA mark Algorithm tagging DKA mark banksta T T T T T P ⁺ T T ⁺ Caudal anterior cingulate F F F F F F F F Caudal middle frontal F F F F F F F F Wedge O O O O O O O O Entorial T T T T T T T T frontal pole F F F F F F F F Fusiform T T T T T T T T Inferior parietal P P P P P P P P Inferior temporal T T T T T T T T Island F F F F F F F F Isthmus of the cingulate gyrus F F F F F F F F Lateral occipital O O O O O O O O Lateral orbitofrontal F F F F F F F F lingual O O O O O O O O Medium orbito-frontal F F F F F F F F Middle temporal T T T T T T T T Paracentral P P P P F ⁺ F ⁺ P ⁺ P ⁺ parahippocampal T T T T T T T T tire part F F F F F F F F Orbital part F F F F F F F F triangular part F F F F F F F F Pericalkarin O O O O O O O O Postcentral P P P P P P P P Back waist P* P* F* F* P P ⁺ P F ⁺ Precentral F* F* P* P* P P P P Fore cunee P P P P P P P P Rostral anterior cingulate F F F F F F F F Rostral-median frontal F F F F F F F F Upper frontal F F F F F F F F Superior parietal P P P P P P P P Superior temporal T T T T T T T T Supramarginal P P P P P P P P temporal pole T T T T T T T T Temporal transverse T T T T T T T T

Abbreviations: F - frontal, T - temporal, P - parietal, O - occipital, L - left, R - right.

Table A.3 - Mean bark thickness (CT) and total surface area (SA) averaged over the four lobes in each study group

Frontal Parietal temporal Occipital CT (mm) average (co) HE 2.56(0.09) 2.67(0.02) 2.22(0.02) 2.05(0.03) bvFTD 2.25(0.06) 2.42(0.08) 2.06(0.02) 1.94(0.01) AD 2.31(0.08) 2.38(0.02) 1.96(0.01) 1.82(0.02) SA (×10 ⁵ mm ² ) medium (co) HE 43.4(0.5) 28.9(0.6) 22.2(0.9) 12.3(0.8) bvFTD 41.6(0.5) 27.5(0.8) 21.6(0.9) 12.0(0.9) AD 41.6(0.5) 27.2(0.9) 21.5(0.9) 12.1(0.9)

Abbreviations: HE - Healthy Elderly Subjects, bvFTD - Behavioral Variant of Frontotemporal Dementia, AD - Alzheimer's Disease

Table A.4 CT network nuclei of frontal, temporal, parietal, and occipital modular organization in HE, bvFTD, and AD. The nuclei were ranked according to their interlobar participation index (p) and intralobar z-score (z). High p/high z indicate so-called integration regions (i.e. nodes that communicate across all lobes) and low p/high z regions are so-called peripheral cores (i.e. nodes that communicate within their own module/ shares).

a) Positive subnet cores

List of nodes with high p/high z

HE Share bvFTD Share AD Share Caudal middle frontal F Caudal middle frontal F Rostral middle frontal F Upper frontal F tire part F Inferior temporal T Superior parietal P Rostral middle frontal F Superior parietal P Pericalkarin O Upper frontal F Wedge O Wedge O Caudal middle frontal F Caudal middle frontal F Upper frontal F tire part F Middle temporal T Middle temporal T triangular part F Superior temporal T Superior parietal P Rostral middle frontal F Superior parietal P Wedge O Upper frontal F Wedge O lingual O List of low p/high z nodes HE Share bvFTD Share AD Share Middle temporal T Fusiform T Lateral orbito-frontal F Superior temporal T Rostral middle frontal F Rostral middle frontal F Inferior temporal T

b) Negative subnet cores

List of nodes with high p/high z

HE Share bvFTD Share AD Share Rostral anterior cingulate F Island F Caudal middle frontal F Middle temporal T Precentral F Island F Transverse temporal T banksta T Transverse temporal T Transverse temporal T Rostral middle frontal F Island F Caudal middle frontal F parahippocampal T Entorhinal T Lateral orbito-frontal F temporal pole T temporal pole T temporal pole T

Abbreviations: HE - healthy elderly subjects, bvFTD - behavioral variant of frontotemporal dementia, AD - Alzheimer's disease, F - frontal, T - temporal, P - parietal, O - occipital

Table A.5 - SA network nuclei of frontal, temporal, parietal and occipital modular organization in HE, bvFTD and AD. The nuclei were ranked according to their interlobar participation index (p) and intralobar z-score (z). High p/high z indicate so-called integration regions (i.e., nodes that communicate across all lobes) and low p/high z regions are so-called peripheral cores (i.e., nodes that communicate within their own module/ shares).

a) Positive subnet cores

List of nodes with high p/high z

HE share bvFTD share AD share Island F Orbital part F Island F Inferior temporal T Lateral orbito-frontal F Middle temporal T Medium orbito-frontal F Orbital part F Rostral anterior cingulate F Middle temporal T Fusiform T Postcentral P Fore cunee P Superior parietal P Middle temporal T Lateral orbito-frontal F temporal pole T banksta T Fore cunee P Middle temporal T Pericalkarin O Pericalkarin O List of low p/high z nodes HE share bvFTD share AD share Lateral orbito-frontal F Orbital part F Orbital part F Inferior temporal T Inferior temporal T Middle temporal T Fore cunee P temporal pole T Pericalkarin O Fore cunee P Lateral orbito-frontal F Middle temporal T Fore cunee P Pericalkarin O Pericalkarin O

b) Negative subnet cores

List of nodes with high p/high z

HE share bvFTD share AD share tire part F Caudal anterior cingulate F frontal pole F Fore cunee P Rostral middle frontal F Lateral orbito-frontal F banksta T Precentral F Transverse temporal T Rostral anterior cingulate F Posterior parietal P tire part T Paracentral P Transverse temporal T Upper frontal F Caudal anterior cingulate F Lateral orbito-frontal F Rostral middle frontal F Upper frontal F

Abbreviations: HE - healthy elderly subjects, bvFTD - behavioral variant of frontotemporal dementia, AD - Alzheimer's disease, F - frontal, T - temporal, P - parietal, O - occipital.

Table A.6 Core networks of CT-SA combination of frontal, temporal, parietal, and occipital modular organization in HE, bvFTD, and AD. The nuclei were ranked according to their interlobar participation index (p) and intralobar z-score (z). High p/high z indicate so-called integration areas (i.e. nodes interacting across all shares)

a) Positive subnet cores

List of nodes with high p/high z

HE share bvFTD share AD share Caudal middle frontal F tire part F temporal pole T Precentral F Transverse temporal T Upper frontal F Supramarginal P Posterior parietal P Posterior parietal P Transverse temporal T Posterior parietal P Caudal middle frontal F Transverse temporal T tire part F triangular part F Upper frontal F Postcentral P Supramarginal P Lateral occipital O

Abbreviations: HE - healthy elderly subjects, bvFTD - behavioral variant of frontotemporal dementia, AD - Alzheimer's disease, F - frontal, T - temporal, P - parietal, O - occipital.

Links

Gauthier, S. et al. “Efficacy and safety of tau-aggregation inhibitor therapy in patients with mild or moderate Alzheimer's disease: a randomised, controlled, double-blind, parallel-arm, phase 3 trial”, The Lancet 388, 2873-2884 (2016)

Wilcock, GK et al “Potential of low dose leuco-methylthioninium bis (hydromethanesulphonate) (lmtm) monotherapy for treatment of mild Alzheimer's disease: Cohort analysis as modified primary outcome in a phase iii clinical trial. Journal of Alzheimer's disease 61, 635-657 (2018)

Feldman, H. et al “A phase 3 trial of the tau and tdp-43 aggregation inhibitor, leuco-methylthioninium bis (hydromethanesulfonate) (lmtm), for behavioral variant frontotemporal dementia (bvFTD)” Journal of Neurochemistry 138, 255 (2016)

Murray, AD et al “The balance between cognitive reserve and brain imaging biomarkers of cerebrovascular and Alzheimer's diseases” Brain 134, 3687-3696 (2011)

Storey, JD “A direct approach to false discovery rates” Journal of the Royal Statistical Society: Series B (Statistical Methodology) 64, 479-498 (2002)

Van Wijk, BC, Stam, CJ & Daffertshofer, A. “Comparing brain networks of different size and connectivity density using graph theory.” PLoS One 5 e13701 (2010)

Rubinov M, Sporns O. “Weight-conserving characterization of complex functional brain networks”, Neuroimage , 56(4):2068-79 (2011)

All links mentioned above are included here by reference.

Claims (71)

1. A method for determining the responses of patients with tauopathy to neuropharmacological intervention in the form of a pharmaceutical product with a symptomatic or disease-modifying effect, including the steps: obtaining first structural neurological data from patients prior to neuropharmacological intervention, wherein these first structural neurological data characterize the physical structure of cortical regions, the physical structure being cortical thickness (CT) and/or cortical surface area (SA); creating the first correlation matrix based on the first structural neurological data by: appointment of structural nodes corresponding to the cortical areas of the brain; And determining pairwise correlations between pairs of structural nodes based at least in part on the respective data of the first structural neurological data; obtaining second structural neurological data from said patients after the neuropharmacological intervention, said second structural neurological data characterizing said physical structure of the cortical regions; And creating a second correlation matrix based on the second structural neurological data by: determining pairwise correlations between pairs of structural nodes based at least in part on the respective data of the second structural neurological data; wherein the method includes: comparing the first correlation matrix and the second correlation matrix and thereby determining the patient's response to the neuropharmacological intervention, wherein the comparison of the first correlation matrix and the second correlation matrix comprises: comparing the number and density of inverse correlations in the first correlation matrix, respectively, with the number and density of inverse correlations in the second correlation matrix, or comparing the number of inverse correlations in the first correlation matrix with the number of inverse correlations in the second correlation matrix, or comparing the density of inverse correlations in the first correlation matrix with the density of inverse correlations in the second correlation matrix. 2. The method of claim 1, wherein a p-value is determined for each pairwise correlation and compared to a significance level, where only p-values less than the significance level are used to generate the corresponding correlation matrix. 3. The method according to claim 1 or 2, where the assignment of structural nodes corresponding to the cortical regions of the brain further includes the definition of groups that contain structural nodes corresponding to the same lobes or different lobes. 4. The method of claim 3, wherein the comparison of the first correlation matrix and the second correlation matrix comprises: comparison of the number and density of correlations, respectively, with the number and density of correlations between different groups of structural nodes corresponding to the same shares, or comparison of the number of correlations with the number of correlations between different groups of structural nodes corresponding to the same shares, or comparing the density of correlations with the density of correlations between different groups of structural nodes corresponding to the same shares, or comparison of the number and density of correlations, respectively, with the number and density of correlations between different groups of structural nodes corresponding to different shares, or comparing the number of correlations with the number of correlations between different groups of structural nodes corresponding to different proportions, or comparison of the density of correlations with the density of correlations between different groups of structural nodes corresponding to different shares. 5. The method according to claim 3 or 4, where the comparison of the first correlation matrix and the second correlation matrix includes: comparison of the number and density of correlations, respectively, with the number and density of correlations between groups of structural nodes located in the frontal lobe and groups of structural nodes located in the parietal and occipital lobes, or comparison of the number of correlations with the number of correlations between groups of structural nodes located in the frontal lobe and groups of structural nodes located in the parietal and occipital lobes, or comparison of the density of correlations with the density of correlations between groups of structural nodes located in the frontal lobe and groups of structural nodes located in the parietal and occipital lobes. 6. The method of any one of the preceding claims, wherein the neuropharmacological intervention is a disease-modifying pharmaceutical. 7. The method according to any one of paragraphs. 1-6, where the neuropharmacological intervention is symptomatic treatment. 8. The method of any one of the preceding claims, wherein the neuropharmacological intervention is a disease-modifying pharmaceutical and efficacy is determined through: reducing the number and density of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix, or reducing the number of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix, or reducing the density of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix. 9. The method according to any one of the preceding claims, wherein the first structural neurological data is obtained by magnetic resonance imaging. 10. A system for determining the responses of patients with tauopathy to neuropharmacological intervention in the form of a pharmaceutical product with a symptomatic or disease-modifying effect, including: a data acquisition means configured to obtain first structural neurological data from patients prior to the neuropharmacological intervention, the first structural neurological data characterizing the physical structure of cortical regions, the physical structure being cortical thickness (CT) and/or cortical surface area (SA); a correlation matrix generator configured to generate a first correlation matrix based on the first structural neurological data via: the appointment of structural units corresponding to the cortical areas of the brain; And determining pairwise correlations between pairs of structural nodes based at least in part on the respective data of the first structural neurological data; wherein the data collection means is also configured to obtain second structural neurological data from said patients after the neuropharmacological intervention, said second structural neurological data characterizing said physical structure of the cortical regions; And the correlation matrix generating means is also configured to generate a second correlation matrix based on the second structural neurological data via: determining pairwise correlations between pairs of structural nodes based at least in part on the respective data of the second structural neurological data; the system additionally includes either: display means for representing the first correlation matrix and the second correlation matrix; or comparison means for comparing the first correlation matrix and the second correlation matrix and thereby determining the patient's response to the neuropharmacological intervention, wherein the comparator is configured to compare the first correlation matrix and the second correlation matrix by: comparing the number and density of inverse correlations in the first correlation matrix, respectively, with the number and density of inverse correlations in the second correlation matrix, or comparing the number of inverse correlations in the first correlation matrix with the number of inverse correlations in the second correlation matrix, or comparing the density of inverse correlations in the first correlation matrix with the density of inverse correlations in the second correlation matrix. 11. The system of claim 10, wherein the system further comprises a tester configured to determine a p-value for each pairwise correlation and compare the p-value to a significance level, wherein the correlation matrix generator is configured to use only p-values less than the significance level when creating a correlation matrix. 12. The system according to claim 10 or 11, where the assignment of structural nodes corresponding to the cortical regions of the brain further includes the definition of groups that contain structural nodes corresponding to the same lobes or different lobes. 13. The system of claim 12, wherein the comparator is configured to compare the first correlation matrix and the second correlation matrix via: comparison of the number and density of correlations, respectively, with the number and density of correlations between different groups of structural nodes corresponding to the same shares, or comparison of the number of correlations with the number of correlations between different groups of structural nodes corresponding to the same shares, or comparing the density of correlations with the density of correlations between different groups of structural nodes corresponding to the same shares, or comparison of the number and density of correlations, respectively, with the number and density of correlations between different groups of structural nodes corresponding to different shares, or comparing the number of correlations with the number of correlations between different groups of structural nodes corresponding to different proportions, or comparison of the density of correlations with the density of correlations between different groups of structural nodes corresponding to different shares. 14. The system of claim 12 or 13, wherein the comparator is configured to compare the first correlation matrix and the second correlation matrix via: comparing the number and density of correlations, respectively, with the number and density of correlations between one or more groups of structural nodes located in the frontal lobe and one or more groups of structural nodes located in the parietal and occipital lobes, or comparing the number of correlations with the number of correlations between one or more groups of structural nodes located in the frontal lobe and one or more groups of structural nodes located in the parietal and occipital lobes, or comparing the density of correlations with the density of correlations between one or more groups of structural nodes located in the frontal lobe and one or more groups of structural nodes located in the parietal and occipital lobes. 15. The system according to any one of paragraphs. 10-14, where the neuropharmacological intervention is a disease-modifying pharmaceutical. 16. The system according to any one of paragraphs. 10-14, where the neuropharmacological intervention is symptomatic treatment. 17. The system according to any one of paragraphs. 10-16, where the neuropharmacological intervention is a disease-modifying pharmaceutical and efficacy is established through: reducing the number and density of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix, or reducing the number of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix, or reducing the density of correlations between the anterior and posterior regions of the brain in the first correlation matrix and the second correlation matrix. 18. The system according to any one of paragraphs. 10-16, where the first structural neurological data is obtained by magnetic resonance imaging. 19. A computer-readable medium that stores a computer program containing executable code that, when executed on a computer, causes the computer to perform the method according to any one of paragraphs. 1-9. 20. The method according to any one of paragraphs. 1-9, or a system according to any one of paragraphs. 10-18, or the computer-readable medium of claim 19, wherein the patient's response to neuropharmacological intervention is determined in the context of a clinical trial to evaluate the effectiveness of a pharmaceutical in the treatment of tauopathy, and the effectiveness of the pharmaceutical is evaluated in whole or in part based on a comparison of the response of a group of patients with a comparison group, who did not receive such intervention.

Images