GB2555853A

GB2555853A - A computer-implemented method for assessing concentration of a subject, and a computer, a system and a computer program therefor

Info

Publication number: GB2555853A
Application number: GB1619222.1A
Authority: GB
Inventors: Townsend Joseph
Original assignee: Fujitsu Ltd
Current assignee: Fujitsu Ltd
Priority date: 2016-11-14
Filing date: 2016-11-14
Publication date: 2018-05-16

Abstract

A computer-implemented method of assessing concentration of a subject, and system and computer therefore, comprises tracking of the subject's eye gaze S10 when reading an assessment text to give one or more numerical results S40 based on either (i) a total return distance S20 provided by calculating the distance by which the subject's eye gaze has returned to a previous word in the assessment text each time the subject's eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance and/or (ii) and a return rate S30 provided by dividing a number of times the subjects eye gaze has returned to a previous position by a time taken to read the assessment text; before outputting the one or more numerical results S50 to assess the subject's concentration. The results of the method can be used in diagnosis of a mental health condition such as attention deficit hyperactivity disorder (ADHD) or depression.

Description

(54) Title of the Invention: A computer-implemented method for assessing concentration of a subject, and a computer, a system and a computer program therefor

Abstract Title: Computer implemented method and system for assessing the concentration of a subject (57) A computer-implemented method of assessing concentration of a subject, and system and computer therefore, comprises tracking of the subject's eye gaze S10 when reading an assessment text to give one or more numerical results S40 based on either (i) a total return distance S20 provided by calculating the distance by which the subject's eye gaze has returned to a previous word in the assessment text each time the subject's eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance and/or (ii) and a return rate S30 provided by dividing a number of times the subjects eye gaze has returned to a previous position by a time taken to read the assessment text; before outputting the one or more numerical results S50 to assess the subject's concentration. The results of the method can be used in diagnosis of a mental health condition such as attention deficit hyperactivity disorder (ADHD) or depression.

1/10

Figure 1

2/10

Figure 2

3/10

Figure 3

4/10

Figure 4

5/10

ο .ο

LD

Φ ι_

Ζ3 .00

LU

§ c:

I cj

V £

JO §

§ s

ju

Return distance; 3 words

LT)

Φ i_ □

.00

LX.

6/10

Start \j oo <υ

u.

□ ω

ο

LT) rd

IZ1

r\

X c

Lk

7/10

QJ

CTi ω

s_ =3

CUO pH cn 00 Ps ft ft ft o o © φ m q· * « ft © o o

PO PM pH © • ft ft ο o ©

B4O3S

8/10

	o	co	LO	CD	CO	CO	co T““	22	24	27
i—	o	o	o	O	o	O	o	o	o	o
o	o	o	o			CM	CM	CM	co	co
Ό	o	o	o			00	co	CO	o	o
Previous word i		Φ JZ 1-	fox	Φ JZ -1—·	brown	over	quick	jumps	lazy	over
lndex(previous word i)	1	o	co	co	CM	LO	V“		r--	LO
Current word i	The	X £	the	brown	over	quick	jumps	lazy	over	dog
lndex(current word i)	o	co	co	CM	LO	V-	M~	h-	LO	co
	o		CM	CO	M·	LO	CO	h-	oo	CD

Figure 10

9/10

ο	σ> LO θ	θ CO	co un co	oo co	co co co
, .		CO	co CD c5	LO co co’	-^r co co
	CM σ> co	CM LO co		CM co	0.44
το	χ—	CM Γ'— CO	oo oo co	co co co	LO LO co
r- £Ζ .2 go CD 3 Ο co ό ΣΕ α_		Γ~- CO CO		co co co	Γ'- co co
•ύέ CD £		CM	CO	'St	LO

	42	30	39	LO	27
		CM	CM	CM
	co	CO	LO	LO	V“
	CM	T—	CM	O	T—
>-	d	d	d	d	d
TO	C3O	co	co	co	o
PHQ-9 Question 7	co	CM	co	T““	CM
Week		CM	co	xf	LO

10/10

Figure 13

9JO3S

A computer-implemented method for assessing concentration of a subject, and a computer, a system and a computer program therefor

The present invention relates to assistance in diagnosis and understanding of a condition, for a human individual or subject. The subject may be a patient of a medical institution, such as a hospital, or clinician’s practice such as a doctor’s or psychotherapist’s practice. The term medical institution can also cover a medical conglomeration across different locations and hospital/practices and also refers to alternative medicine practices. Alternatively, the subject may be an employee, or potential employee, child or student or any other human individual.

In many scenarios, a subject may already be suffering from a disorder, but in others the subject is currently healthy. The invention may be applied in diagnosis/objective assessment of poor concentration as well as mental conditions such as ADHD (Attention Deficit Hyperactivity Disorder), and other mental or physical disorders, illnesses and diseases, such as autism, anxiety and depression. The invention is thus widely applicable in provision of healthcare and in assessment of concentration for other purposes, such as educational, skill profiling purposes or comparison of individuals.

Subjects being tested for depression or other health issues, such as ADHD, may be asked to complete a questionnaire on a regular basis, in which they self-assess the frequency of various problems, including their inability to concentrate. One problem with this self-diagnostic test is that it is subjective in nature and therefore may not necessarily reflect the subject’s concentration ability as accurately as it could. If concentration could be quantified more objectively at the same (or greater) frequency than the questionnaire is taken, this would, for example, support a doctor or therapist in forming a diagnosis or at least in questioning the accuracy of a subjects’ self-diagnosis. Moreover, even if there is no known mental health condition or even any suspicion of one, an objective concentration test could be applied on a routine basis.

In particular, subjects being tested for depression are sometimes asked to complete a

Patient Health (PHQ) Questionnaire, in which they self-assess the frequency of nine problems over the past two-week period, assigning scores corresponding to frequency.

Figure 1 shows an example Patient Health Questionnaire (PHQ-9), taken by patients (or potential patients) in the UK for the diagnosis of depression. Similar questionnaires are used elsewhere for the same purpose. The subject self-assesses the frequency of each of nine problems, assigning a score corresponding to frequency. The total ‘PHQ-9’ score across the nine questions is calculated and used to determine severity of the depression. A similar questionnaire, GAD-7, exists in the UK for the diagnosis of generalized anxiety disorder (GAD). Again, this is the test taken in the UK and similar tests in other countries are equally relevant examples. As a subject is treated for any problems identified, therapists or doctors may ask the subject to retake the questionnaire on a regular basis in order to monitor progress (or lack thereof).

The Patient Health Questionnaire, and other self-assessed health questionnaires, are subjective. Objective methods for automatically completing health questionnaires (but not specifically the PHQ-9) have been proposed. These methods are based on readings of, for example, physical activity data, heart rate data, etc. Detection of stress levels in speech has also been proposed as a means of diagnosis for stress-related mental health problems. It is desirable to assist the medical practitioner in interpreting self-assessed health questionnaires and other tests, particularly those which relate to mental heaith. It is also desirable in general to find a way of objectively assessing concentration levels.

Statements of Invention

According to an embodiment of a first aspect of the invention, there is provided a computer-implemented method of assessing concentration of a subject comprising: assessing the subject’s concentration by tracking of the subject’s eye gaze when reading an assessment text to give one or more numerical results; wherein the one or more numerical results are based on at least one or more of the following factors: a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a said previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a said previous position by a time taken to read the assessment text; and outputting the one or more numerical results to assess the subject’s concentration.

The inventor has come to the realisation that one problem with the Patient Health Questionnaire PHQ-9 and other self-diagnosis tests for mental or general health, or even for poor concentration, is that they are subjective in nature. Different subjects may interpret the questions or symptoms differently or struggle to remember the frequency of their own problems, especially if the very nature of their mental health problem hinders their judgement. Furthermore, subjects in denial of health or concentration problems may not answer honestly. If the answer to a question relating to diagnosis could be quantified more objectively at the same (or greater) frequency than the frequency at which a questionnaire or other self-assessment test is taken, this would support the doctor, therapist, teacher or other examiner in forming a diagnosis or at least questioning the accuracy of the subjects’ answers.

While objective auto-completion methods for health questionnaires do exist, none target questions which appear in tests such as the PHQ-9 questionnaire specifically, and many of the questions in these or other mental health tests are, by their very nature, perhaps so subjective as to be difficult to quantify. For example, although an answer to question 3 in the PHQ-9 questionnaire, which concerns sleep, could be quantified using existing methods for sleep monitoring, questions such as 1 and 2 (‘little interest or pleasure’ and ‘feeling down’ respectively) are mental symptoms which cannot be measured using existing sensor-based methods. Lack of concentration (question 7) is one of these questions for which no method of quantification or auto-completion has been proposed.

While a machine learning approach may help with diagnosis, existing machine learning approaches to gaze-based mental health diagnosis require a vast amount of past data from other subjects, which would require storage and processing resources, and may raise security and privacy concerns.

Invention embodiments explain how a solution may be found based on reading ability as measured using gaze-tracking technology. Reading speed is traditionally based on words, lines or pages per minute, and it may be tempting to think that a quick reading speed suggests that the reader is assimilating information quickly and thus is concentrating. However the inventor has come to the appreciation that this (at least on its own) is not an accurate measure of concentration. The reader may move between words quickly but return to previously read text that he or she may be struggling to understand.

Embodiments of the invention may provide a method of quantifying a lack of concentration based on positional data collected from gaze-tracking technology. For example, each time the subject takes their depression questionnaire, they may be asked to complete a reading exercise in which gaze-tracking technology tracks what they are reading on the page. Based on the assumption that when a subject is struggling to concentrate they may pause or back-track to re-read segments of text, the invention embodiments record instances of these behaviours to produce metrics which reflect the severity of the subject’s inability to concentrate.

The method may include repeating the concentration assessment on a periodic basis; and outputting the one or more numerical results for each of the periodic concentration assessments. This may allow an appreciation of changes in concentration assessment over time, and potentially generation of an alert if concentration deteriorates over time. Equally, if one or more assessments give a result below a threshold, demonstrating an unacceptable or dangerously low level of concentration, an alert may be generated.

The gaze-tracking data used in the invention may detect fixation of the eye on words, characters, or even sentences, and use this data to calculate a return distance.

For example, the subject’s concentration may be assessed by tracking of the subject’s eye gaze on words when reading the assessment text. This eye gaze on a word may be on a single character of a word, or on more than one character per word, to recognise the word. Also, in some languages a word is formed from a single written character. Calculating the return distance may be achieved by counting the number of words by which the subject’s eye gaze has returned to a previous word in the assessment text each time the subject’s eye gaze has returned to a previous word. Equally, providing the return rate may be by dividing a number of times the subject’s eye gaze has returned to a previous word by a time taken to read the assessment text.

Alternatively, the software may consider fixation of the eye on single characters, consider the position of the eye in terms of characters and use this data to calculate a return distance expressed in terms of characters.

In a third alternative, the software may consider fixation of the eye on a sentence as a whole (perhaps by allocating each fixation to the sentence in which the fixation is positioned), and process the position of the eye in terms of sentences, using this data to calculate a return distance expressed in terms of sentences.

The metrics which reflect the severity of the subject’s inability to concentrate, whether expressed in terms of words, character or sentences, may be plotted against a selfassessment score reflecting the subject’s own assessment, so that an examiner (such as a teacher or medical practitioner) may observe whether or not a correlation exists, which would suggest whether or not the subject provided an accurate self-assessment.

Hence the method may further comprise the subject rating their concentration level in a self-assessment questionnaire at one or more times when the concentration assessment takes place. The self-assessed concentration level and the one or more numerical results may be output (for example to an examiner) for comparison. This allows direct comparison of the objective results with subjective results.

The method may further use a correlation metric (any know correlation mechanism) to assess the correlation between the one or more numerical results and the self-assessed concentration level (for example over one or more instances when the self-assessment questionnaire and test are taken on the same occasions) and demonstrate accuracy of the subject’s self-assessment.

Accordingly, the method may further comprise: comparing the self-assessed concentration level and the one or more numerical results and generating an alert indicating poor seif-assessment when the self-assessed concentration level and the one or more numerical results do not match to a required extent. Any alert may be provided as a sound alert and/or on a display or printer and/or sent (to an examiner) over a network, for example by email. It may allow a dangerous discrepancy in the subject’s self-assessment and their objective condition to be identified automatically.

Additionally or alternatively, when the concentration assessment and self-assessment are repeated on a periodic basis, the method may further comprise: comparing the selfassessed concentration level and the one or more numerical results and generating an alert indicating decreasing self-assessment ability when the self-assessed concentration level and the one or more numerical results deviate over time. This may allow a dangerous deterioration in self-assessment to be identified automatically.

The method may further comprise: normalising the one or more numerical results and the self-assessed concentration level. This allows direct comparison. Any suitable normalisation method may be used to give a range of outcomes between 0 and 1, or in any other suitable range. When the concentration assessment and self-assessment are repeated on a periodic basis, the method may then also comprise outputting the normalised numerical results and the normalised self-assessed concentration levels as a graph over time of the normalised numerical results and self-assessed concentration level.

In some embodiments, the one or more numerical results may also be based on a factor of the time taken to read the assessment text.

Each factor (for example the time taken, the return count and the total return distance) may be output as an individual (potentially normalised) numerical result for each concentration assessment.

Alternatively or additionally, the method may include averaging the factors on which the numerical results are based. Hence one output may be a lack of concentration result as a single (potentially normalised) numerical result for each concentration assessment. The factors may be weighted before averaging.

According to an embodiment of a second aspect of the invention, there is provided a system to assess concentration of a subject comprising: an eye-gaze device to track the subject’s eye gaze when reading an assessment text; and a computer to: control the eye-gaze device, to read in the eye-gaze data and to give one or more numerical results for each assessment; wherein the one or more numerical results are based on at least one or more of the following factors: a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and to output the one or more numerical results.

The computer may control the eye-gaze device by instructing what to monitor, for example fixation on words or on characters, or other parameters.

According to an embodiment of a third aspect of the invention, there is provided a computer to assess concentration of a subject comprising: an input to receive eye-gaze data from an eye-gaze device tracking the subject’s eye gaze when reading an assessment text; a processor to read in the eye-gaze data and give one or more numerical results for each assessment; wherein the one or more numerical results are based on at least one or more of the following factors: a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and one or more of a display and a network interface to output the one or more numerical results to assess the subject’s concentration.

According to an embodiment of a fourth aspect of the invention, there is provided a computer program, which when executed on a computer provides a method of assessing concentration of a subject comprising: assessing the subject’s concentration by controlling tracking of the subject’s eye gaze when reading an assessment text to give one or more numerical results; wherein the one or more numerical results are based on at least one or more of the following factors: a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and outputting the one or more numerical results to assess the subject’s concentration.

An apparatus or computer program according to preferred embodiments of the present invention may comprise any combination of the method aspects. Methods or computer programs according to further embodiments may be described as computerimplemented in that they require processing and memory capability.

The apparatus according to preferred embodiments is described as configured or arranged to, or simply “to” carry out certain functions. This configuration or arrangement could be by use of hardware or middleware or any other suitable system. In preferred embodiments, the configuration or arrangement is by software.

Thus according to one aspect there is provided a program which, when loaded onto at least one computer configures the computer to become the apparatus according to any of the preceding apparatus definitions or any combination thereof.

According to a further aspect there is provided a program which when loaded onto the at least one computer configures the at least one computer to carry out the method steps according to any of the preceding method definitions or any combination thereof.

In general the computer may comprise the elements listed as being configured or arranged to provide the functions defined. For example this computer may include memory, processing, and a network interface.

The invention may be implemented in digital electronic circuitry, or in computer hardware, firmware, software, or in combinations of them. The invention may be implemented as a computer program or computer program product, i.e., a computer program tangibly embodied in a non-transitory information carrier, e.g., in a machine-readable storage device, or in a propagated signal, for execution by, or to control the operation of, one or more hardware modules.

A computer program may be in the form of a stand-alone program, a computer program portion or more than one computer program and may be written in any form of programming language, including compiled or interpreted languages, and it may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a data processing environment. A computer program may be deployed to be executed on one module or on multiple modules at one site or distributed across multiple sites and interconnected by a communication network.

Method steps of the invention may be performed by one or more programmable processors executing a computer program to perform functions of the invention by operating on input data and generating output. Apparatus of the invention may be implemented as programmed hardware or as special purpose logic circuitry, including e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for executing instructions coupled to one or more memory devices for storing instructions and data.

The invention is described in terms of particular embodiments. Other embodiments are within the scope of the following claims. For example, the steps of the invention may be performed in a different order and still achieve desirable results. Multiple test script versions may be edited and invoked as a unit without using object-oriented programming technology; for example, the elements of a script object may be organized in a structured database or a file system, and the operations described as being performed by the script object may be performed by a test control program.

Elements of the invention have been described using the terms “input”, “processor”, “output” etc. The skilled person will appreciate that such functional terms and their equivalents may refer to parts of the system that are spatially separate but combine to serve the function defined. Equally, the same physical parts of the system may provide two or more of the functions defined.

For example, separately defined means may be implemented using the same memory and/or processor as appropriate.

Preferred features of the present invention will now be described, purely by way of example, with references to the accompanying drawings, in which:Figure 1 shows a Patient Health Questionnaire (PHQ-9), taken by subjects for the diagnosis of depression, in which the subject self-assesses the frequency of each of nine problems, assigning a score corresponding to frequency;

Figure 2 is a flow diagram of a general embodiment of the invention;

Figure 3 is a hardware diagram of a system and a computer according to a general invention embodiment;

Figure 4 is a flow diagram of one embodiment of the overall examination process;

Figure 5 shows an assessment text recognised by an eye-gaze device, in which the current word is ‘over’ (word number 5)

Figure 6 shows an assessment text recognised by an eye-gaze device, in which the new current word is ‘dog’ (word 8), and the previous word, i.e. the word on which the subject was previously fixated, is Over’ (word number 5);

Figure 7 shows an assessment text recognised by an eye-gaze device, in which the new current word is ‘brown’ (word 2), and the previous word is ‘over’ (word number 5);

Figure 8 is a flowchart of a data collection process according to an invention embodiment;

Figure 9 is a plot of a weekly PHQ-9 score against metrics m = [d, r, t] and lack of concentration C as a function of m;

Figure 10 is a table of variable changes as the subject reads the text “The quick brown fox jumps over the lazy dog” on week 5 of their treatment for depression;

Figure 11 is a table showing scores for all five weeks of testing in the worked example; Figure 12 is a table showing normalised score for all five weeks of testing in the worked example, complete with concentration score C, based on Equation 3 with weights w = [0.7, 0.5, 0.6] for m = [d, r, t]; and

Figure 13 is a plot of all the scores in Figure 12.

General Description

The use of ‘gaze tracking’ (also known as ‘eye tracking’) for medical diagnosis is known. ‘Gaze-tracking’ refers to the act of detecting movement in the eyes and/or where an individual is looking, i.e. what exactly their eyes are concentrating on. Numerous technologies for tracking gaze exist, one such being the SMI Eye Tracking Glasses (http://www.eyetracking-glasses.com/). In the prior art, gaze-tracking may be used for the diagnosis of problems in the eye itself, but also has been suggested for the diagnosis of mental health problems by reading gaze data from a patient and classifying a diagnosis by matching that data to other patients of the same diagnosis using a trained machine-learning algorithm, see US2016/106315, for example.

Gaze-tracking has also been suggested as a means of diagnosing dyslexia, as the technology has been widely used to measure reading speed. Gaze tracking is used to identify fixations, i.e. characters or words in the text on which the reader is focussed, so that the rate at which the reader moves between words can be calculated or estimated. The reader does not necessarily fix his gaze on every word in order to read as they may read unfocussed words in their peripheral vision. Normally reading speed is based on words, lines or pages per minute, though in the dyslexia diagnosis application the diagnosis may be based on the duration of a fixation on individual words.

Invention embodiments use gaze tracking in a different way to provide a solution to the problem of automatically quantifying concentration levels when a subject takes a health assessment test, such as the PHQ-9 test. Each time the subject takes the test, they may be asked to complete a reading exercise while wearing gaze-tracking apparatus which follows which words or characters they are focussed on. Invention embodiments assume that taking longer to read the text or back-tracking to earlier segments of text suggests that the subject is struggling to concentrate, and quantify lack of concentration by quantifying instances of these events. Three metrics may be recorded: the total ‘return distance’ (the total ‘distance’ crossed (for example in word, character or sentence count) when a subject returns to an earlier segment of text), the total number of times the subject returns to an earlier segment of text (regardless of distance), and the total time taken to read the document.

For example, each time the subject takes the PHQ-9 questionnaire and the concentration exercise, their score for question 7 on the questionnaire and each metric may be plotted on a graph, along with the same values recorded for previous examinations. This enables the examiner to observe whether or not the subject’s self-assessment of their own concentration on the questionnaire correlates with the objective metrics recorded during the reading exercise, and therefore how accurate their self-assessment may be.

While a machine learning approach may help with diagnosis, it requires a vast amount of historical data from other subjects. Some invention embodiments may use the realisation that it is not so much quantifying lack of concentration (or any other symptom) that is important, but quantifying it with respect to the previous weeks’ (or other periodic instances of) scores in order to assess how well it correlates with the subjects’ selfassessment. Hence invention embodiments may require only the historic data of the subject in question and not that of other subjects. This reduces requirements for storage and processing resources, and maintains privacy of subjects not concerned with this diagnosis.

Figure 2 is a diagram showing a process for assessing concentration of a subject. Firstly, S10 tracks the subject’s eye gaze when reading an assessment text. This tracking can allow calculation of a total return distance provided by calculating the distance by which the subject’s eye gaze has returned (moved back) to a position previously read (for example a previous word, character or sentence in the reading direction) each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach a total return distance (S20). Additionally or alternatively, this tracking can allow calculation of a return rate provided by dividing a number of times the subject’s eye gaze has returned to a word, character or sentence previously read by a time taken to read the assessment text (S30).

One or both of these factors are used to give one or more numerical results in S40, for example by providing both factors and a numerical result for each, or by averaging the factors to give a single numerical result.

The optional periodic repeat loop shown in dashed lines indicates that S10 to S40 may be repeated as required to give an assessment over time, for example over a given period of time (for example weekly or fortnightly) or for a given number of iterations.

Finally, S50 outputs the one or more numerical results, for each of the periodic concentration assessments if the repeat loop is included. For periodic assessments, this may take place incrementally after each assessment or after any repeats have finished.

Figure 3 is a block diagram of a system 10, which comprises a computing device 30 and an eye-gaze device 20. The computing device, such as a data storage server, embodies the present invention, and, together with the eye gaze device to implement the physical tracking, may be used to implement a method of an embodiment of the invention to assess subject concentration and to carry out the method shown in figure 2. The computing device comprises a processor 993, and memory, 994. Optionally, the computing device also includes a network interface 997 for communication with other computing devices, for example with other computing devices of invention embodiments, or to send results of the assessment over a network.

For example, an embodiment may be composed of a network of such computing devices. The computing device also includes one or more input mechanisms including an input mechanism for eye-gaze data from the eye-gaze device 20, and possible others such as keyboard and mouse 996, and a display unit such as one or more monitors 995. The components are connectable to one another via a bus 992.

The memory 994 may include a computer readable medium, which term may refer to a single medium or multiple media (e.g., a centralized or distributed database and/or associated caches and servers) configured to carry computer-executable instructions or have data structures stored thereon. Computer-executable instructions may include, for example, instructions and data accessible by and causing a general purpose computer, special purpose computer, or special purpose processing device (e.g., one or more processors) to perform one or more functions or operations.

Thus, the term “computer-readable storage medium” may also include any medium that is capable of storing, encoding or carrying a set of instructions for execution by the machine and that cause the machine to perform any one or more of the methods of the present disclosure, such as the methods described in figures 2, 4 and 6 and the methods defined in the claims. The term “computer-readable storage medium may accordingly be taken to include, but not be limited to, solid-state memories, optical media and magnetic media. By way of example, and not limitation, such computer-readable media may include non-transitory computer-readable storage media, including Random Access Memory (RAM), Read-Only Memory (ROM), Electrically Erasable Programmable ReadOnly Memory (EEPROM), Compact Disc Read-Only Memory (CD-ROM) or other optical disk storage, magnetic disk storage or other magnetic storage devices, flash memory devices (e.g., solid state memory devices).

The processor 993 is configured to control the computing device and execute processing operations, for example executing code stored in the memory to implement the various different functions described here and in the claims. For example, the processor can execute code to calculate the total return distance and the return rate during an assessment, as well as any further desired metrics. The memory 994 stores data being read and written by the processor 993. As referred to herein, a processor may include one or more general-purpose processing devices such as a microprocessor, central processing unit, or the like. The processor may include a complex instruction set computing (CISC) microprocessor, reduced instruction set computing (RISC) microprocessor, very long instruction word (VLIW) microprocessor, or a processor implementing other instruction sets or processors implementing a combination of instruction sets. The processor may also include one or more special-purpose processing devices such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), network processor, or the like. In one or more embodiments, a processor is configured to execute instructions for performing the operations and steps discussed herein.

The display unit 997 may display a representation of data stored by the computing device and may also display a cursor and dialog boxes and screens enabling interaction between a user and the programs and data stored on the computing device. Furthermore, the display unit may display tables and graphs of the assessment results to the user and display any alert created. Additionally or alternatively, the tables and graphs (and any alert) may be forward via an output, such as over the network interface. The input mechanisms 996 may enable a user to input data and instructions to the computing device. For example, the user may input a preferred set of metrics to be used in assessing concentration.

The network interface (network l/F) 997 may be connected to a network, such as the Internet, and is connectable to other such computing devices via the network. The network l/F 997 may control data input/output from/to other apparatus via the network. Other peripheral devices such as microphone, speakers, printer, power supply unit, fan, case, scanner, trackerball etc. may be included in the computing device.

Methods embodying the present invention may be carried out on a computing device such as that illustrated in Figure 3. Such a computing device need not have every component illustrated in Figure 3, and may be composed of a subset of those components. A method embodying the present invention may be carried out by a single computing device in communication with one or more data storage servers via a network. The computing device may be a data storage itself storing the assessment data.

A method embodying the present invention may be carried out by a plurality of computing devices operating in cooperation with one another. One or more of the plurality of computing devices may be a data storage server storing at least a portion of the assessment results.

Detailed Embodiment

As part of their treatment for depression, a patient is asked to take the PHQ-9 questionnaire (Figure 1) on a regular basis.

Each time the patient takes the questionnaire, they are also asked to complete a reading exercise to quantify their concentration levels while being observed by gaze tracking hardware and/or software. This exercise supports their self-assessment in answering question 7 of the questionnaire.

For a fair examination, the reading text should ideally be different from week to week since the patient may find it easier to concentrate on something they have read before. However since there can only be a finite number of text samples, the text used furthest in the past should be used when all other samples have been used. For a fair examination all text samples should be of equal or similar length and difficulty.

Figure 4 is a flow diagram of one embodiment of the overall examination process. In S60, the patient takes the PHQ-9 questionnaire. The score for question 7 is added to the results. At the same time (at least approximately, for example on the same day) the patient completes the reading exercise in S70. Any of the metrics/factors described herein are used to quantify lack of concentration in S80 and a score is added to the results. In S90, the results are plotted. In S100 the examiner makes an assessment or diagnosis.

Figures 5 to 7 show a sentence, with a dashed circle indicating a current word on which a subject (such as the patient referred to above) is fixated, and a dotted circle indicating the previous word on which the subject was fixated. Each word is associated with an index which identifies its position in the text sequence. As the subject moves from word to word, the apparatus keeps track of two words (and their corresponding indices); the current word and the previous word. It is important to note that when during reading, the eye gaze does not necessarily make a smooth sweep across words but moves in a series of short stops, or ‘fixations’. The ‘current word’ refers to the word on which the subject is currently fixated, and the ‘previous word’ refers to the word on which the subject was previously fixated, which is not necessarily earlier in the text since the reader may backtrack on what they have already read.

The gaze-tracking apparatus (for example the SMI Eye Tracking Glasses mentioned previously) is used to identify where the subject’s gaze is fixed, and therefore which word in the text they are fixed on.

In figure 5, the apparatus detects just the current word on which the subject’s gaze is fixated. As mentioned previously, all the words are assigned an index. In this figure, the current word is Over’ (word number 5).

Figure 6 is an example of when the subject moves forward in the text when reading. The new current word is ‘dog’ (word 8), and the previous word, i.e. the word on which the subject was previously fixated, is ‘over’ (word number 5).

Figure 7 is an example of when the subject moves backwards in the text when reading. The new current word is ‘brown’ (word 2), and the previous word is ‘over’ (word number 5). When the subject moves backwards in the text, a ‘return distance’ is measured as the difference between the two indices (Equation 1 below).

Of course, this is a much- simplified example and the actual assessment text or texts will be considerably longer.

Certain reading behaviours suggest poor concentration. If the subject is not concentrating, they may not assimilate information properly as they read, and therefore may need to return to earlier parts of the text more often in order to do so. Furthermore, if the subject becomes distracted or struggles to assimilate information, they may also pause on a word for an extended period of time. Therefore, in order to detect poor concentration levels, the setup is used to identify when the subject returns to earlier text segments or pauses on words.

As the user reads, three metrics are recorded: total return distance d (Equation 1), return rate r (Equation 2) and total reading time t.

Return distance measures how far (with respect to word count) the user returns to previous segments of text which they have already read, and return rate measures how often they do this per second. Total reading time will be longer if the subject pauses on words more often or for longer periods of time.

Return distance d i _ V^-1 (index(previous word i) — (current word t) if index(current word) < index(previous word t) ^_ Z_i I 0 if index(current word i) > (previous word i) n=l

Equation 1: Total return distance d.

number of returns to an earlier word return rate r = —---:—-— ---— time taken to read whole text (seconds)

Equation 2: Return rate r

The process for recording these metrics is summarised in Figure 8, which is a flowchart of the data collection process. Return rate r and return count are used here. Thus for scoring purposes, if the fixation moves forward in the text (Figure 6), no further action is taken. If the fixation moves backwards in the text (Figure 7), the return count c is incremented, and movement distance with respect to the number of words is added to the total return distance d. A timer records t, the total time in seconds the subject spends reading, by starting the timer when the subject reads the first word and stopping it when they read the last. After the last word has been read, return rate r is calculated by dividing return count c by time t. The process is described in more detail below.

In S110, the reading timer is started. In S120 it is determined whether the fixation of the reader on a word has moved.

If the fixation has moved in S120, there is a decision in S130 as to whether it has moved forwards or backwards. If it has moved backwards, the return count c is incremented by 1 in S140 and the return distance d is incremented by the number of words moved backwards in S150. The process then continues at S120.

If the fixation has moved forwards in S120, it is determined whether the reader has reached the end of the text in S160: if so, the timer stops in S170 and the return rate is calculated by dividing the return count by the reading time in S180. If the reader has not reached the end of the text in S160, the process continues at S120.

Other metrics may be recorded as the subject reads the text, if the examiner feels they are useful in diagnosing a poor level of concentration, for example reading speed in words per minute. Any metrics recorded may be recorded in a vector m. The examiner may also choose to exclude d, r and t from m if they so wish.

Each time the subject takes the PHQ-9 questionnaire and the concentration exercise, their score for question 7 on the questionnaire and each metric are plotted on a graph, along with the same values recorded for previous examinations. A sample plot can be seen in Figure 9, plotting weekly PHQ-9 score again metrics m = [d, r, t] and lack of concentration C as a function of m. All of the points are normalised by the maximum for each value (or by any other means) before they are plotted. A function on the scores for each metric may also be plotted. An estimate for lack of concentration C is calculated for example as shown by Equation 3 below using the set of metrics m and corresponding weights w where w' is the weight in the range [0,1] associated with metric m'. m may include any combination of d, r and t and any other metrics the examiner may wish to include. Values for w may be selected by the examiner or by any other means.

_{c =} n

Equation 1: Lack of concentration C

The plots generated over time will inform the examiner as to how well the subjects’ selfassessment correlates with empirical observations of their concentrations as observed by the reading tests, and therefore how accurate the subjects’ self-assessment is likely to be. Correlation metrics may be used to assess this more empirically. The correlation is not necessarily intended to make the diagnosis on behalf of the examiner, but to support the examiner in making their own diagnosis.

The results plot and examiner’s assessment may also be used with metrics for quantifying any other question in the PHQ-9 questionnaire, or any other questionnaire, mental health related or otherwise. Data collection for quantifying answers may be performed by any equipment, apparatus or test that is suitable for the question being answered.

Worked Example

A subject has been treated for depression over a five week period, taking both the PHQ9 questionnaire (Figure 1) and the reading exercise each week.

It is now week 5, and the subject answers ‘2’ for question 7, to say that they have been struggling to concentrate for more than half the days over the past two weeks.

The subject is then given the task of reading the text “The quick brown fox jumps over the lazy dog” (Figures 5 to 7) while wearing gaze-tracking apparatus, for example the SMI Eye Tracking Glasses . As the subject reads the text, the following takes place. Figure 10 shows the variable changes as the subject reads the text “The quick brown fox jumps over the lazy dog” on week 5 of their treatment for depression.

Taken line by line, Figure 10 shows the following

0. The subject reads word 0 (‘The’) and the timer t starts.

1. The subject’s gaze moves to word 3 (‘fox’)

2. The subject’s gaze moves to word 6 (‘the’).

3. The subject’s gaze then moves back to word 2 (‘brown’). Because the gaze has returned to a previous segment of text, the return count c is incremented by 1, and the total return distance d is incremented by 4, since the difference between index of the current and previous words is 4 (i.e. ‘brown’ and ‘the’ are 4 words apart).

4. The subject’s gaze moves to word 5 (‘over’), d and c remain unchanged since the gaze is moving forward again.

5. The subject’s gaze moves back again, to word 1 (‘quick’). The return count c is incremented to 2, and the total return distance d is incremented to 8.

6. The subject’s gaze moves forward to word 4 (’jumps’).

7. The subject’s gaze moves forward to word 7 (‘lazy’).

8. The subject’s gaze moves back one last time to word 5 (‘over’). The return count c is incremented to 3, and the total return distance d is incremented to 10.

9. The subject’s gaze moves forward to word 8 (‘dog’). As this is the final word, the timer stops at 27 seconds. The return rate r can now be calculated according to Equation 2, by dividing return count c by reading time t, i.e. 3127 ~ 0.111.

The final scores for week 5 can be compared to the same over the previous four weeks. This is shown by the table in figure 11. They are then normalised so that an estimate of the level of lack of concentration C (Equation 3) can be calculated (as shown in figure 12).

The normalised scores, along with the calculated concentration levels are then plotted on a graph so the examiner can look for potential correlations (see figure 13).

All scores generated from the data appear to correlate with the subject’s answers to question 7 of the PHQ-9 questionnaire, suggesting that the subject’s self-assessment of their concentration levels is indeed accurate.

If there is no such correlation, or if the scores are low, or if the correlation decreases over time an alert may be produced, to warn the examiner of the discrepancy.

Benefits

Invention embodiments may provide a novel approach to quantifying the concentration ability of an individual based on a score which quantifies how often that individual returns to earlier segments of text when reading, and how this score compares to readings of the same score in previous weeks.

Three novel quantification methods may be used:

• A measure of total ‘return distance’ in word count when the user’s gaze returns to a point of fixation earlier in the text;

• A measure of the rate at which the user’s gaze returns to a point of fixation earlier in the text by dividing the number of such returns by the total time taken to read the text.

• A weighted function which takes either or both of the above as inputs, along with any other metrics the examiner may wish to include.

The approach to quantifying concentration may be applied to the diagnosis of any illness 10 where poor concentration is believed to be a symptom, for example ADHD.

The metric plotting aspect of invention embodiments is applicable to any questionnaire where one or more of its questions can be quantified based on data collected autonomously. However the method of collection and quantification may need to be specific to each question.

Invention embodiments may support doctors and therapists in making more accurate diagnoses for their subjects and/or assessing the accuracy of a subject’s self-diagnosis. With better diagnoses, the subjects will be able to receive better treatment.

Invention embodiments require only the data of the subject being diagnosed, unlike existing machine learning approaches to medical diagnosis using gaze-tracking, which require historical data from multiple subjects. With minimal data being used for diagnosis, processing and storage requirements are lessened, and data security and privacy concerns may be alleviated.

Claims

1. A computer-implemented method of assessing concentration of a subject comprising:

assessing the subject’s concentration by tracking of the subject’s eye gaze when reading an assessment text to give one or more numerical results; wherein the one or more numerical results are based on at least one or more of the following factors:

a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and outputting the one or more numerical results to assess the subject’s concentration.

2. A method according to claim 1, including repeating the concentration assessment on a periodic basis; wherein the outputting includes:

outputting the one or more numerical results for each of the periodic concentration assessments.

3. A method according to claim 1 or 2, including:

assessing the subject’s concentration by controlling tracking of the subject’s eye gaze on words when reading the assessment text;

calculating the return distance by counting the number of words by which the subject’s eye gaze has returned to a previous word in the assessment text each time the subject’s eye gaze has returned to a previous word; and providing the return rate by dividing a number of times the subject’s eye gaze has returned to a previous word by a time taken to read the assessment text.

4. A method according to any of the preceding claims, further comprising:

the subject rating their concentration level in a self-assessment questionnaire at one or more times when the concentration assessment takes place; and outputting the self-assessed concentration level and the one or more numerical results for comparison.

5. A method according to claim 4, further comprising:

using a correlation metric to assess the correlation between the one or more numerical results and the self-assessed concentration level and demonstrate accuracy of the subject’s self-assessment.

6. A method according to claim 4 or 5, further comprising:

comparing the self-assessed concentration level and the one or more numerical results and generating an alert indicating poor self-assessment when the self-assessed concentration level and the one or more numerical results do not match to a required extent.

7. A method according to claim 4, 5 or 6 further comprising:

repeating the concentration assessment and self-assessment on a periodic basis; comparing the self-assessed concentration level and the one or more numerical results and generating an alert indicating decreasing self-assessment ability when the self-assessed concentration level and the one or more numerical results deviate over time.

8. A method according to any of claims 3 to 6 further comprising:

normalising the one or more numerical results and the self-assessed concentration level;

repeating the concentration assessment and self-assessment on a periodic basis; and outputting the normalised numerical results and the normalised self-assessed concentration levels as a graph over time of the normalised numerical results and selfassessed concentration level.

9. A method according to any of the preceding claims, wherein:

the one or more numerical results are also based on a factor of the time taken to read the assessment text.

10. A method according to any of the preceding claims, comprising: outputting each factor as an individual numerical result.

11. A method according to any of the preceding claims, comprising:

averaging the factors on which the numerical results are based to output a lack of concentration result as a single numerical result.

12. A method according to claim 11, comprising:

weighting the factors before averaging the factors to provide the lack of concentration result.

13. A system to assess concentration of a subject comprising:

an eye-gaze device to track the subject’s eye gaze when reading an assessment text;

a computer to:

control the eye-gaze device, to read in the eye-gaze data and to give one or more numerical results for each assessment; wherein the one or more numerical results are based on at least one or more of the following factors:

a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous word by a time taken to read the assessment text; and to output the one or more numerical results to assess the subject’s concentration.

14. A computer to assess concentration of a subject comprising:

an input to receive eye-gaze data from an eye-gaze device tracking the subject’s eye gaze when reading an assessment text; and a processor to read in the eye-gaze data and give one or more numerical results for each assessment; wherein

5 the one or more numerical results are based on at least one or more of the following factors:

a total return distance provided by calculating the distance by which the subject’s eye gaze has returned to a previous position in the assessment text each time the subject’s eye gaze has returned to a previous position, to form a return distance, and by

10 summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and one or more of a display and a network interface to output the one or more numerical results to assess the subject’s concentration.

15. A computer program, which when executed on a computer provides a method of assessing concentration of a subject comprising:

assessing the subject’s concentration by controlling tracking of the subject’s eye gaze when reading an assessment text to give one or more numerical results; wherein

20 the one or more numerical results are based on at least one or more of the following factors:

25 summing any return distances to reach the total return distance; and a return rate provided by dividing a number of times the subject’s eye gaze has returned to a previous position by a time taken to read the assessment text; and outputting the one or more numerical results to assess the subject’s concentration.

Intellectual

Property

Office

Application No: Claims searched:

GB 1619222.1 1 to 15