WO2012065215A1 - Electrical stimulation for the suppression of epileptic seizures - Google Patents

Abstract

A device and method for counteracting seizure events in a mammalian brain is described in which a plurality of electrodes are configured to deliver pulsatile electrical stimulus to the brain. The pulses are applied across the electrodes asynchronously. At each electrode, the pulses can be periodic or aperiodic. The pulse width may be greater than 300 μsec

Description

P

1

"Electrical stimulation for the suppression of epileptic seizures"

Cross-Reference to Related Applications

The present application claims priority from Australian Provisional -Patent Application No 2010905090 filed on 16 November 2010, the content of which is incorporated herein by reference.

Field

The present invention relates to use of an electrical stimulus of specific characteristic, applied by multiple electrodes placed within or near the source or sources of seizures in the brain, to terminate or at least counteract epileptic seizure events arising from those sources.

Background

Epilepsy is a chronic neurological disorder characterised by recurring unprovoked seizures, which are symptoms of episodic abnormal electrical activity in the brain. Epilepsy is a group of syndromes with widely varying symptoms, but all involving abnormal neuronal activity. Epilepsy is quite common, with about 0.5 - 1 % of the population having active epilepsy at any time.

Generalized epilepsy is produced by electrical activity that arises substantially simultaneously throughout the entire brain. In contrast partial seizures are produced, at least initially, by electrical impulses that are generated in a relatively small part of the brain, referred to as the focus.

Medicines exist that can assist in controlling epileptic seizures in some patients. For drug refractory epilepsy patients, efforts are made to locate the focus causing the seizures with a view to surgical removal of a lesion or the like at that location, if possible. However, only a small proportion of drug refractory epilepsy patients can be- treated surgically. To provide another treatment option, recent work has used implanted electrical devices to apply electrical stimuli. Under a number of different stimulation approaches, the stimuli have been applied to either the vagus nerve in the neck, to the brain surface, or to the focus within the patient's brain. Responsive stimulation involves monitoring electrical activity of the brain to detect signs that a seizure is imminent or has commenced, and applying the stimulus only in response to such detection. The applied stimulus is intended to suppress the seizure.

As seen in electroencephalography (EEG), normal ongoing brain activity is mostly in the frequency range of lHz (sleep) to 40Hz (gamma range, active

concentration), and is predominantly between lHz and 20Hz, Epileptiform electrical activity can be abnormally synchronous, and one previous approach applies electrical stimulus configured to desynchronise such electrical activity.

The stimulus can be applied via one or more electrodes. In conventional apparatus, when a plurality of electrodes are employed, the pulses of the electrical stimulus are applied across the electrodes synchronously or within a time window that is short relative to the time between successive pulses on the same electrode (pseudo- synchronously). In the latter case, due to the nature of signal generators used, pulses may be delivered to each electrode sequentially, but the time delay between delivery of . the pulses is intentionally very short so that the pulses occur within a short time window and the duty cycle for stimulation is small. Conventionally, the electrodes are stimulated in the same fixed sequence.

Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application.

Throughout this specification the word "comprise", or variations such as

"comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

Summary

According to a first aspect, the present invention provides a method for counteracting seizure events in a mammalian brain, the method comprising applying electrical stimulus to the brain via a plurality of electrodes, the electrical stimulus I

3 comprising a train of electrical pulses, the application of the pulses across the electrodes being asynchronous.

According to a second aspect, the present invention provides a device for counteracting seizure events in a mammalian brain, the device comprising:

a plurality of electrodes for delivering an electrical stimulus to the brain; and a stimulus generator which is arranged to apply pulsatile electrical stimulus to each of the electrodes such that pulses are applied across the electrodes

asynchronously.

By intentionally applying the pulses across the plurality of electrodes asynchronously, i.e., not consistently at substantially the same time across the electrodes, rather than synchronously or pseudo-synchronously, some of the adverse effects of electrical stimulation can be avoided or reduced. It has been found, for example, that the epileptiform afterdischarge duration (EAD), and seizure severity, of a mammalian patient suffering a seizure event can be reduced significantly when asynchronous stimulation is carried out in accordance with the present invention.

In the first and second aspects the pulses may be applied to the electrodes in sequential order or a variable order, e.g. a randomised variable order, and the stimulus generator may be considered to "cycle" through the plurality of electrodes in that order. Whether the pulses are applied in sequential order or in varying order, pulses may be applied to electrodes at times substantially across an entire cycle period. This may occur in e.g. all or the majority of the cycles, and may be such that at least one of the pulses applied within the cycle period occurs after e.g. 50%, or 60% or 75% of the cycle period, and/or so that there is no gap in the cycle period in which no pulses are applied of greater than e.g. 25%, 40%, 50%, 60% or 75% of the cycle period. When the electrodes are addressed in a variable order, the cycle period may be the period in which all electrodes are addressed at least once, prior to one of the electrodes being addressed again. This approach, where pulses are applied to electrodes at times substantially across an entire cycle period, may contrast with a pseudo-synchronous approach, where a group of pulses across a group of electrodes may occur in a fixed relationship to one another and in a fixed order within a short time period, e.g. less than P

4

500 ms, for example, and then there may be a gap of several seconds before the next group of pulses.

The pattern of pulses in a pulse train at each electrode may be the same.

Alternatively one or more characteristics of the pulse train may differ between different electrodes. For example, although the inter-stimulus interval (ISI) in the pulse train at each electrode may be consistent, the ISI may be different between different electrodes, providing one example of periodic asynchronous stimulation. As another example, the ISI may vary at one or more, e.g. all, of the electrodes, for example, by independent Poisson processes, whether at the same average pulse rate or at distinct average pulse rates, providing aperiodic asynchronous stimulation. On the whole, the pulses applied to the brain, may be periodic (i.e., with constant ISI) or aperiodic (i.e. with varying ISI), and the pulse rate, or average pulse rate, may be the same, or different, at each electrode. When the pulses are aperiodic, there may be no clear order of pulses across all electrodes at least over a short time period. More than one pulse may be applied at one electrode before a pulse is applied to another electrode, for example. Nonetheless, over an extended time period, the number of pulses applied at each electrode may average out.

According to a third aspect, the present invention provides a method for counteracting seizure events in a mammalian brain, the method comprising applying electrical stimulus to the brain via a plurality of electrodes, the electrical stimulus comprising a train of electrical pulses, the pattern of the pulses applied by each electrode being different.

According to a fourth aspect, the present invention provides a device for counteracting seizure events in a mammalian brain, the device comprising:

a plurality of electrodes for delivering an electrical stimulus to the brain; and a stimulus generator which is arranged to apply pulsatile electrical stimulus to each of the electrodes such that the pattern of pulses applied at each electrode is different.

The first to fourth aspects recognise that repetitive predictable patterns of stimulation in time and space may be more effective at triggering seizures, and therefore less effective at terminating seizures. It is considered that disruption of an ]

5 ordered excitation of the neurons in the brain may be achieved by using a disordered series of electrical pulses in time and space, at least over relatively short time periods. In accordance with the present, invention, a disordered series of electrical pulses in time and space may be achieved through asynchronous and optionally aperiodic application of electrical pulses via the electrodes, for example. The stimulus applied across the plurality of electrodes may, over an extended period of time, have a temporal and spatial pattern that is substantially uniform. However, the stimulus may have a low level of predictability from one pulse to the next.

In certain embodiments, to ensure charge balance, a plurality of electrode pairs may be provided. Although stimuli may be applied substantially synchronously at the two electrodes of each electrode pair (in bipolar fashion), pulses across different electrode pairs are nonetheless applied asynchronously. In general, reference in this application to stimuli applied at different electrodes may include stimuli applied across different single electrodes (in monopolar fashion) or different electrode pairs (in bipolar fashion).

Preferably at least 4 electrodes are employed in the asynchronous stimulation. For example 5 or more electrodes may be employed or 10 or more electrodes or 16 or more electrodes may be employed. In one embodiment, 28 electrodes are employed and in other embodiment 120 electrodes are employed.

In any of the aspects, using multiple electrodes can allow a degree of randomness in the stimulation to be achieved both in time and space. Furthermore, it can enable a larger or smaller stimulus current to be applied in a more controlled fashion and for stimulus to be applied to a larger or smaller area of the brain in a controlled fashion. The use of multiple electrodes with asynchronous and optionally aperiodic stimulation, as described above, potentially lowers the stimulus current required at each electrode to achieve a therapeutic effect, in comparison to a single electrode with a fixed rate stimulus, for example.

When the pulses are periodically applied at each electrode, the stimulus generator may cycle through the electrodes in turn or in a variable order, with the time delay between successive pulses being substantially proportional to the total time period for each cycle, divided by the number of electrodes (n). Thus, the I

6 asynchronicity can be maximised in comparison to synchronous or pseudo-synchronous approaches. When the pulses are applied aperiodically at each electrode, there may be no obvious sequence or cycle of pulses applied across the electrodes at least over a short period of time. The application of pulses at each electrode and/or across all of the. electrodes may be randomised.

The mean frequency of stimulation may be between 1 and 300Hz. For example, the mean frequency of stimulation may be 60Hz. In alternative embodiments, the mean frequency of stimulation may be greater than 300Hz.

In any of the aspects of the present invention, the electrical stimulus may be of very short duration, for example comprising fewer than 10 pulses. Alternatively, the electrical stimulus may be applied substantially continuously over an extended period. The electrical stimulus may preferably be applied to the brain for between 300ms and 3000ms. Pulses within the pulse train may all have the same amplitude, or may differ in amplitude. Pulses within the pulse train may all have the same pulse width, or may differ in pulse width. Different sections of the pulse train may have different pulse rates or different Poisson distributed pulse intervals.

Some embodiments of the invention may further comprise monitoring electrical activity of the brain to detect a seizure event, and applying the electrical stimulus only in response to a seizure event being detected. For example the device may comprise a signal processor configured to analyse electrical activity of the brain detected by one or more sensors. One or more electrodes may serve as sensor electrodes(s). In such embodiments of the invention, the stimulus is preferably initiated earlier than two seconds after onset of a seizure. In embodiments of the invention, the seizure event may comprise an impending seizure or an occurring seizure. Prediction of the impending seizure or detection of occurrence of a seizure may be performed by any suitable method.

Alternative embodiments may apply the electrical stimulus substantially continuously or at regular intervals, without attempting to detect a seizure event.

In embodiments of the invention, the sensor of the device may comprise one or more electrodes of the plurality of electrodes. Thus, the electrode(s) used or usable to P(

7 apply the stimulus may also function as sensor electrode(s). Alternatively, separate electrodes may be used as sensor electrodes.

In embodiments of the invention, the stimulus may be adaptable in response to measured brain electrical activity before, during or after application of one or more preceding stimuli. Alternatively, in some embodiments the stimulus to be applied may be fixed, with brain activity detection used only for detection of a seizure event.

Where the patient in question is known to have focal seizures, the stimulus is preferably applied to or near the focus of the seizure activity within the brain.

Embodiments of the device of the present invention may comprise:

an implantable component comprising the electrodes and an internal antenna; and

an external component comprising a power source, the processor, and an external antenna for wireless transcutaneous communication with the internal antenna.

The internal antenna and external antenna preferably comprise inductive coils for effecting inductive transcutaneous power transmission to power the implantable component and optionally transfer data in both directions for the detection of seizures and control of stimuli. In alternative embodiments, wired communication may be use in place of wireless communication.

As indicated, the device may comprise a number of connected or integrated parts and/or a number of separate, distinct parts. The device may also be considered an apparatus.

The mammalian brain may comprise a human brain.

The stimulus generator may be any means for applying pulsatile electrical stimuli to the electrodes in accordance with the desired stimulation paradigms, e.g. in an asynchronous and optionally aperiodic manner, and with certain frequencies and/or pulse widths. In one embodiment, the stimulus generator may comprise a switch matrix configured to individually address the electrodes and a power source connected to the switch matrix.

The average frequency of stimulation applied to the brain by the combination of the plurality of electrodes may be greater than 1 Hz or greater than 1 OHz or greater than 50Hz or higher. The frequency may be greater than 300Hz in some embodiments. It is 1*^|

8 recognised that application of a stimulus to counteract a seizure event can be efficacious when a series of pulses are delivered to the brain at a frequency substantially higher than frequency ranges previously considered clinically relevant to electrical activity in the brain. In some embodiments, the average frequency may greater than 320 Hz, greater than 350 Hz, or greater than 370 Hz. In some embodiments, the average frequency may be less than 5 kHz, less than 2.5 kHz, less than 1 kHz. less than 700 Hz, less than 600 Hz, or less than 500 Hz. In one embodiment, the average frequency may be substantially 370 Hz. However, it has been recognised herein that the efficacy of seizure termination may be increased using asynchronous pulsatile stimulation across multiple electrodes, and optionally with pulse durations of e.g. greater than 300 μ8εο, substantially independently of the frequency of stimulation.

The pulse duration (pulse width) applied by each electrode may be greater than 300 μεεο in some embodiments. For example, the pulse width may be in the range of 300-1000 μββϋ or 300-1500μ8εϋ.

Brief Description of the Drawings

Examples of the invention will now be described with reference to the accompanying drawings, in which:

Figure 1 is a schematic drawing of apparatus according to one embodiment of the present invention;

Figure 2 is a top view of an implantable component electrode array for use in the apparatus of Figure 1 ;

Figure 3 is a schematic representation of pulse timings for four stimulation modes applied to the brain via four bipolar, stimulating electrode pairs (pairs 1 to 4);

Figure 4 is a schematic representation of two successive biphasic pulses;

Figure 5 is another schematic representation of pulse timings for four stimulation modes applied to the brain via four bipolar, stimulating electrode pairs (pairs 1 to 4);^■ Figure 6 shows histogram plots of stimulating frequencies applied via the four stimulating electrode pairs for each of the four stimulating modes over five non- consecutive stimulations of the same stimulating mode, with mean stimulation frequency for each histogram being shown in the top right hand corner of each plot;

Figure 7 shows a histogram plot of mean epileptiform afterdischarge duration

(in seconds) for each of the four stimulation modes (with *p<0.001 , Mann- Whitney Rank Sum Test), where each error bar indicates one standard error of measurement;

Figure 8 shows a histogram plot of mean (Racine) seizure severity for each of the four stimulation modes (with *p<0.001, Mann- Whitney Rank Sum Test), where each error bar indicates one standard error of measurement.

Figure 9 shows the duration and treatment latency and the total EAD duration for all trials, with and without aperiodic/asynchronous, 500Hz-based, 1000μ8 stimuli;

Figure 10 shows the duration and treatment latency and the total EAD duration for all trials, with and without aperiodic/asynchronous, lOOOHz-based, 500μ5 stimuli;

Figure 11 shows the duration and treatment latency and the total EAD duration for all trials, with and without periodic/synchronous, 125Hz, 300μ8 stimuli;

Figure 12 shows the duration and treatment latency and the total EAD duration for all trials, with and without periodic/synchronous, 125Hz, 1000μ5 stimuli;

Figure 13 shows the duration and treatment latency and the total EAD duration for all trials, with and without aperiodic/asynchronous, 125Hz-based, 1000μ8 stimuli;

Figure 14 shows the duration and treatment latency and the total EAD duration for all trials, with and without periodic/synchronous, 500Hz-based, 1000μ5 stimuli;

Figure 15 shows the duration and treatment latency and the total EAD duration for all trials, with and without aperiodic/asynchronous, 500Hz-based, 1000μ8 stimuli; and

Figure 16 shows a box and whisker plot of seizure durations for different stimulation paradigms.

Description of Embodiments

The present invention arises from consideration of parameters of electrical pulsatile stimulation, such as biphasic rectangular pulsatile stimulation, for seizure J

1 0 abatement. In exploring these potential stimulation parameters, it is required to stay ^•within safety bounds for the charge density (30μΟΌη²).

We first describe apparatus suitable for use in the present invention. The apparatus comprises an implantable component that can take the form of a hermetically sealed, biocompatible metal (e.g. titanium), with polymer or ceramic casing that can be implanted either locally (in the skull) or remotely in the chest cavity or shoulder region of patients if required. The estimated weight of such a device can be less than lOOg and can be of a similar mass to any plate of skull removed during implantation. As the precise brain region to be recorded/stimulated will change from patient to patient, the placement o^'f the device may be standardised to ensure a routine surgical implantation procedure. Skull site may be somewhere around the temporal/occipital region where it is less likely to sustain forceful blows, with the electrode leads crossing the dura in a single location before diverging to specific brain locations. Internal electronics of the implanted component may be responsible for delivery of stimuli and acquisition of neural signals only.

The implantable component may use electrode components suitable for long term monitoring of deep brain or cortical structures, or both. Custom electrode designs may be used to provide higher spatial resolution placement in the areas of interest where standard spacing grid arrays may not suit. For example, 32 individual electrode sites could be implanted in and on the brain, with the intention of only 16 or less of these being monitored or stimulated at any one time. Redundancy of this nature allows for temporal changes in the epilepsy and for failure in particular electrode sites. Power demands of the device may be minimised by ensuring electrode impedances are kept as low as possible for stimulation, which consequently requires high input impedance on the preamplifiers for recording. Surface electrodes may be employed, or deep tissue electrodes.

In one embodiment a "closed loop" approach is taken whereby stimuli are only delivered to the brain tissue in response to detection of a seizure or imminent seizure. Alternative embodiments may adopt an "open loop" approach in which stimuli are delivered without determining whether or not a seizure is occurring or imminent. In the closed loop embodiment, stimuli should be delivered only in response to predicted or J

1 1 detected seizures with a latency as short as possible and preferably at most being 2 seconds or less for detection. Switching times between record and stimulate modes may not be greater than 100μ5 and ideally as short as possible. The combination of electrodes used for stimulation and reference sites can be configurable at the time of implantation and through the device lifetime with the possibility to dynamically alter these sites in response to ECoG cortical signals or other response parameters.

External componentry can be in proximity to the implanted component but externally located. In the external componentry, a processor responsible for communicating with, directing and powering the implanted component may be used. Having the intelligent components outside of the body facilitates upgrades to processor or battery capacity as new technologies become available. Specific processor selection will be dependent on the algorithms requiring implementation. To allow clinicians the best possible information for device assessment the processor should be capable of recording and storing a set amount of data around recognised seizure events. This may be achieved by maintaining a rolling five minute window of data that is used within the detection algorithm, with data older than five minutes being lost. On detecting an event this data can be stored in the onboard memory for future review. If 4 channels are to be stored with 16-bit resolution at 200 Hz sampling rate, the data rate would be 5.6 MB per hour. With 1 GB of memory available within the device, for example, a total of approximately 175 hours of data could be stored for review. For training and refinement purposes the processor should also be able to connect to a laptop and port all incoming data in real-time to an external proprietary capture and visualising program. This connection also allows for reprogramming of the device as individual requirements change and as algorithms for seizure detection and electrical stimulation improve.

Power transfer to the implanted component, and data communication between the componentry, can be effected by magnetic induction and used to transmit detected signals to the external componentry that in turn issues stimulation commands to the internal component in response to algorithm outputs. Power transmitted via the induction loop can render the componentry as relatively simple, thus minimising the chance of failure and the need for subsequent surgical intervention. J

12

Apparatus according to one embodiment of the present invention is shown schematically in Fig. 1. The apparatus includes an implantable component 1 comprising an electrode array 10, which is configured for insertion in and/or on a subject's brain. The electrode array 10 is designed to both monitor for seizure events within the brain and to apply electrical stimulation to the brain via electrodes of the array following detection of a seizure event. In this regard, the electrode array 10 is configured to deliver signals to, and receive signals from, control circuitry.

In this embodiment, the control circuitry includes a switch matrix 1 1 connected to both a stimulator 12 and an EEG acquisition device 13. The control circuitry also includes a processor 2. The switch matrix 1 1 is configured, upon control by a CPU 24 of the processor 2, to route electrical stimulation signals from the stimulator 12 to the electrode array 10, the stimulation being in the form of pulses delivered to a plurality of stimulator electrodes of the electrode array 10 asynchronously. The switch matrix 1 1 is also configured, upon control by the CPU 24, to route seizure detection data from one or more sensor electrodes of the electrode array 10 to the EEG acquisition device 13. A battery 25 is included in the processor 2, along with a communications interface 26 that connects the CPU 24 and acquisition device 13 to a user interface 3, such as a personal computer, e.g., a tablet PC. The user interface 3 may permit the user to monitor seizure detection information and adjust stimulation parameters. Details such as the current stimulation parameters and historical stimulation data may be stored in a digital memory 27 connected to the CPU 24 and the communications interface 26.

The switch matrix 1 1, stimulator 12 and EEG acquisition device 13 may be provided, along with electrode array 10, in the implantable component 1 , as shown in Fig. 1. Alternatively, one or more of the switch matrix 11, stimulator 12 and EEG acquisition device 13 may be located externally to the implantable component 1 and connected to the implantable component 1 via wires.

The communication between the processor 2 and the switch matrix 11 , stimulator 12 and EEG acquisition device 13 may be wireless communication, or may be via wires.

The electrode array 10 for an implantable component 1 according to an embodiment of the invention is shown in more detail in Fig. 2. The array 10 comprises J

13

120 electrodes in this embodiment. The electrodes are divided into two groups.

Electrodes 101 of the first group (represented by the larger circles in Fig. 2) are configured to form bipolar stimulating electrode pairs, and electrodes 102 of the second group (represented by the^' smaller circles in Fig. 1 ) are configured to form bipolar sensing electrode pairs. The electrodes form the stimulating and sensing pairs through their interfacing with a switch matrix 1 and CPU 24, as described above with respect to Fig, 1 , for example, although alternative embodiments may employ hard wiring of the electrodes with stimulating and sensing componentry, or otherwise.

In this embodiment, the electrode array is to be implanted beneath the dura on the surface of the cortex. However, in alternative embodiments, a plurality of depth electrodes may be used which are implanted bilaterally into deep brain structures such as the hippocampus, for example. The electrode arrays may be custom built, or may be pre-manufactured arrays such as grid electrode arrays or multi-strip subdural electrodes, as manufactured by Ad-Tech Medical Instrument CorporationTM, for ^' example.

In use, the electrode array may be positioned adjacent a brain lesion, for example, to counteract epileptic episodes caused by the lesion. Possible relative positioning of the array and a brain lesion and associated interictal activity, is shown in Fig. 2. The relative positioning of the lesion, which may be detected by an MRI scan, is indicated by line 103, and the relative positioning of areas of interictal activity are indicated by lines 104. A traditional approach to cancelling or reducing the effects of the lesion is resection, in the area of the brain indicated by line 105, for example. However, by using electrical stimulation as proposed herein; resection may be unnecessary. By providing a plurality of stimulation electrodes 101 and sensing electrodes 102, over a relatively large grid area, appropriate electrodes can be selected to sense seizure-related activity in the brain and to provide localised stimulation of the brain to counteract seizure events. When the interictal activity is sensed by sensing electrodes 102 in the exemplary regions 104 indicated in Fig. 2, bipolar stimulating electrode pairs 101a, 101b, strategically positioned adjacent the interictal activity, may be selected and used to apply stimulation at targeted areas of the brain. i

14

In this embodiment, the pulses of the electrical stimulus are applied across the plurality of stimulating electrode pairs asynchronously, i.e., not consistently at substantially the same time. Furthermore, the pulses are applied aperiodically, such that, at least over a short period of time, the stimulation pulse pattern at each stimulating electrode pair, and across all of the electrode pairs, is disordered. An example of a stimulation pattern applied to four stimulating electrode pairs is shown in Fig. 3. Each pulse is indicated by a vertical bar 31 , and the sequence of pulses applied to each stimulating electrode pair is represented by the bars 31 of the same horizontal row. The time delay between successive pulses is proportional to the horizontal spacing between vertical bars 31. As can be seen in a box 32, showing an expanded view of the stimulation pattern over a period of 0.1 seconds, the application of pulses at each electrode pair over this time period is substantially random. As indicated, this stimulation approach can be effective at preventing or counteracting seizures as it may disrupt an ordered excitation of the neurons in the brain.

In the example shown in Fig. 3, across the entire 2.0 s stimulation period shown, the mean inter stimulus interval (ISI) at each electrode pair is 16.3 ms. To aid understanding of the term ISI, Fig. 4 shows a schematic drawing of two biphasic stimulation pulses 41 , 42, with the ISI between the two pulses 41, 42 indicated by gap 43. The ISI is the time period between the end of one of the pulse 41 and the start of the next pulse 42. This is different to the inter-phase gap (IPG) which is the time period between the negative phase 41 a, 42a and positive phase 41b, 42b of each pulse, as indicated by gap 44.

The apparatus may disrupt or modify organised neural activity, whether this is of normal or pathological origin. The primary application may be the abatement and elimination of epileptic seizures. It is envisaged that the primary application may be in patients who have intractable epilepsies that are known to be of focal origin on clinical or neurophysiological grounds, but where the structural abnormality responsible is in an area which is either too large or inaccessible to be treated by surgical removal, or located in eloquent brain. The primary target group then are those patients with a discrete cortical focus or foci. However, it is envisaged that a stimulation strategy in accordance with one or more aspects of the invention could also have application for a number of other conditions in which organised neural activity leads to dysfunction, including epilepsies of generalised origin. It is possible that the type of electrical stimulation described here could also have therapeutic benefit for people suffering from a number of neurological disorders including movement disorders such as Parkinson's disease, dystonias, chronic pain, and psychiatric disorders such as depression and obsessive-compulsive disorder. Such embodiments are thus encompassed within the scope of this application, notwithstanding that the primary application considered here is the abatement and elimination of epileptic seizures.

Example 1: Experimental Method and Materials

Implantable components comprising a microwire electrode array and ground and reference screws were surgically implanted in 5 adult male, inbred, Sprague Dawley rats (250-550g).

Each implantable component was a high-impedance, microwire electrode array device, custom built for implantation via a craniotomy opening, and included sixteen 50μιη Tungsten microwire electrodes. The electrodes were arranged in two rows of eight electrodes (i.e., in a 2 x 8 electrode array). Of the sixteen available electrodes, eight were configured as sensing electrodes for electrocorticography (ECoG), and the remaining 8 electrodes were configured as bipolar stimulating electrode pairs for stimulating the brain. In addition to the electrodes, ground and reference electrode wires were provided. The inter-electrode separation in each row was 500μιη and the inter-row separation was ΙΟΟΟμηι, giving an overall coverage of about 4mm x 1mm. The surface area of the active, non-insulated area at each electrode tip was

approximately 4000μη 2 (40 x 10-6cm2). The electrode lengths of the most medial row (when implanted) were 3 mm and the electrode lengths of the most lateral row (when implanted) were 2.5 mm. The difference in length allowed for clearance of the component with the sides of a craniotomy once inserted and allowed for both rows to be inserted to a similar depth when the device was implanted with a slight angle. i

16

Prior to implantation of the implantable component, the rats were placed under general anaesthetic using a combination of Ketamine (75mg/kg i.p.), Xylazine

(lOmg/kg i.p.) and Isoflurane in oxygen. Following induction of general anaesthesia, a cruciform pattern of two incisions was made over the skull midline, each incision 20mm in length in the anteroposterior (AP) and mediolateral (ML) orientations respectively, allowing access to the electrode device implantation site. Hydrogen peroxide was then applied with cotton wool to remove the periosteum and expose and dry the skull to aid in dental cement adhesion. The rats were then located in a stereotaxic headframe to ensure accurate positioning of a craniotomy. The craniotomy opening had a size of 5mm x 3mm in the AP and ML orientations respectively, and with its centre located on the primary motor/somatosensory cortical, M1/S1HL, region of the right hemisphere with AP: 0mm and ML: +1.5mm according to the Paxinos and Watson stereotaxic coordinates scheme (2005).

Using the stereotaxic frame's manipulator arm, the electrode array device was advanced towards the cortical surface until the point of contact and a dorsoventral measurement was taken. The device was then further inserted until the proximal ends of both rows of electrodes had clearly punctured the dura (as verified by visual observation), at which point the direction of travel was reversed until an overall insertion depth of 1mm was achieved. NeuroSeal™ dural sealant (NeuroNexus Technologies, Ann Arbor, MI) was used to fill the void left by the craniotomy and protect the surface of the brain from contact with dental cement applied later in the surgery. Two screw electrodes (1.4 x 3mm) were also secured to the skull via bur holes over the right and left parietal cortices and attached to the microwire electrodes via the two (stainless steel) reference and ground electrode wires, prior to the entire construct being secured with dental acrylic cement (Vertex Dental B.V., Zeist, The Netherlands). The screw electrodes provided ECoG ground and reference connections and contributed to the stability of the headmount construct. Silk sutures were then used to aid in closure of the wound. The rats were then removed from the frame, and placed on a heating pad to aid their recovery from the anaesthesia and were allowed 5 to 7 days recovery prior to cortical stimulation or monitoring/recording. Surgical implantation of the electrodes took approximately 45 minutes per rat. Example 1: Local Field Potential recordings

Following application of a brief inhalational anaesthetic (Isoflurane, 4% in oxygen at 2L/min, Baxter HealthCare Pty Ltd, NSW, Australia) to facilitate cable attachment, intracortical recordings were made in awake, freely moving animals using an acquisition system comprising the combination of a Tucker-Davis Technologies (TDT) RX5 Pentusa Base Station, a RA16PA Medusa Preamp and a RA16CH high impedance headstage, along with acquisition software custom designed for use in the TDT OpenEx environment. Signals from the eight sensor electrodes were acquired at 24.4 kHz, highpass filter 2.2 Hz, lowpass filter 7.5 kHz, before being downsampled and stored at 3.1 kHz. Animals were continuously monitored by an observer during experimental sessions to ensure proper equipment function and that the animals did not become entangled in the attached cables during seizures. ECoG was recorded continuously during experimental sessions, including periods immediately before, during and after electrical stimulation.

Example 1: Electrical Stimulation Methods

The same customised software suite that was used to control the acquisition system was also employed to control timing for stimulus delivery via a TDT RX7 Microstimulator Base Station, an MS 16 Stimulus Isolator and a NC48 Battery Pack.

With reference to Fig. 5, pulsatile electrical stimuli were delivered in a bipolar fashion across the four pairs of stimulating electrodes along the length of the array, in four different modes: periodic/synchronous (PS), aperiodic/synchronous (ApS), periodic/asynchronous (PAs) or aperiodic/asynchronous (ApAs). Regardless of periodicity or synchronicity, the stimulation duration was 2 seconds, the pulse width per phase was set to Ims and the current intensity was 50μΑ per bipolar pair, giving constant charge per phase and charge density values of 50nC and 1273μΟΌη2/ρη35ε respectively.

Periodic stimulation was achieved by sending pulses (triggers) to the relevant electrode addresses in the stimulator once every 16.7ms (60 Hz). For aperiodic stimulation, the mean inter-stimulus interval (mISI) for a given period of stimulation 2011/001473

P

18 needed to equal that of periodic stimulation for the same duration to ensure that, on average, the same net charge was being delivered in both cases. A Poisson process can be used as a simple model of neural firing rates and so was selected here as the method of varying inter-stimulus intervals (ISIs). A Poisson process has probability mass function as set forth in Equations 1 and 2, where k is the number of occurrences of an event in a given period (in this case, the number of pulses delivered in a second) and λ/5_/ is the expected number of events for that same period, i.e. the mean stimulation frequency Ι/μ/sz.

Equation 1

f (k; X_1S] ) = ^ls'^^■— Equation 2

In software this was realised using Equation 3, where rand is any function that generates random numbers evenly distributed in the bounds (0, 1 ].

-„_T - \n{rand)

ISI = - Equation 3

Since each individual stimulus (pulse) consisted of two 1 ms phases, ISIs of duration 2ms or less could result in undesired behaviour from the stimulator, as pulses would have been sent before the completion of the preceding pulse. Consequently, the final distribution of ISIs was hard limited at the lower bound to be greater than 2ms. An upper limit of 38ms served to prevent unreasonably long, yet statistically possible, intervals from being generated. Because of the introduction of these limits, the distribution of ISIs is referred to in this application as band-limited-Poissonian (BLP).

Electrical stimuli were delivered to the four stimulating electrode pairs

(electrode pairs 1 to 4) in four different modes, as illustrated in Figs. 5A to 5D. The pulse timings for the four stimulation modes over the four electrode pairs are illustrated P

19 schematically using vertical bars. Each vertical bar indicates when a single biphasic (lms/phase), 50μΑ pulse was delivered to each stimulating electrode pair.

Referring to Fig. 5A, in one mode, identical pulse trains with constant ISIs at 60Hz frequency, were applied at the same time across all four electrode pairs

(periodic/synchronous stimulation). Referring to Fig. 5B, in another mode, identical pulse trains with constant ISIs at 60Hz frequency, were applied across all four electrode pairs with an n*4.16ms delay, where n - 0, 1 , 2 or 3 for each electrode pair respectively (periodic/asynchronous stimulation). Referring to Fig. 5C, in another mode, identical pulse trains with BLP distributed ISIs at 60 Hz mean frequency (km), were applied at the same time across all four electrode pairs (aperiodic/synchronous stimulation).

Referring to Fig. 5D, in the final mode, different pulse trains with BLP distributed ISIs, and with mean frequency k_/si) of 60 Hz were applied across all four electrode pairs (aperiodic/asynchronous stimulation).

Individual recording/stimulation sessions typically lasted between 1 and 2 hours, including a 10- 15 minute period of habituation of the rats in the test cage at the beginning of the session without electrical stimulation being applied. Stimuli were delivered once every seven minutes, up to a maximum of ten stimuli in any given session. At the session start, two of the four possible stimulation strategies were randomly selected, with each strategy being allocated five of the ten available stimulations. The specific order in which the stimulations were delivered was assigned in a pseudo-randomised fashion.

Example 1: Data Analysis

Epileptiform afterdischarge duration (EAD) and behavioural seizure severity as measured by a modified Racine scale (where 0 = no observable seizure, 1 = chewing and facial clonus, 2 = head nodding, 3 = forelimb clonus, 4 = rearing, and 5 = rearing and falling) were used to grade the tolerability of the varying stimulus types. Using the sensor electrodes, ECoG data was accumulated and processed and seizure times were measured via visual analysis of recorded ECoG .traces by an independent viewer beginning immediately after the cessation of stimulation artifacts. Racine measures were evaluated at the time of data collection and, where required, verified using videos recorded during the experiment. ^"Both EAD duration and Racine seizure severity measures were expressed as mean ± one standard error of measurement (SEM).

SigmaStat™ (SPSS Inc., Illinois, USA) was used to implement nonparametric Mann- Whitney Rank Sum tests and significance for all tests was set at pO.001.

Example 1: Results

In order to ensure that stimuli were being delivered in the fashion intended, the stimulation pulse times were recorded as epochs for each bipolar stimulating electrode pair, at each stimulation time. Figure 6 shows histogram plots of stimulating frequency for each of the four stimulating modes for each stimulating electrode pair, where the plots in column A are for the periodic/synchronous (PS) mode; column B are for the periodic/asynchronous (PAs) mode; column C are for the aperiodic/synchronous (ApS) mode and column D are for the aperiodic/asynchronous (ApAs) mode. Data for each mode was collected from the stored epochs of five stimuli of the same mode type that were delivered during a single, randomly selected stimulating session. Stimulating frequency was determined by calculating the inverse of each of the ISIs recorded in successive stimuli.

Histograms for the PS and PAs stimuli in columns A and B show that 100% of the delivered stimuli were at a single stimulating frequency (60Hz) for all electrode pairs. The mean stimulation frequencies shown in the top right hand corners of the plots confirm that only periodic stimuli of 60Hz were presented during these events. On the other hand, the histograms for ApS stimuli in column C highlight the presence of the BLP distribution of ISIs, since a spread of stimulation frequencies between 26Hz and 500Hz is confirmed. The identical morphologies of the histograms across the electrode pairs, and the identical mean stimulation frequencies in the top right hand comer of the plots, also serve to validate the synchronous nature of the stimuli delivered across the four pairs. Histograms in column D represent the distributions of the ApAs stimuli. In these plots the BLP and individual distributions, and the different mean stimulating frequencies of the histograms for each electrode pair, confirm both the aperiodic and asynchronous nature of the delivered stimuli. ]

21

Results for the EAD durations and seizure severities that were recorded are shown in Figs. 7 and 8 respectively. Statistical significance (pO.001) was reached for comparisons between any synchronous and asynchronous stimulation types for both duration and severity results, regardless of the periodicity of stimulation. No significant statistical differences were seen in comparisons of stimulation types with the same synchronicity attributes for either duration or severity.

Referring to Figs. 7 and 8, it can be seen from the histogram heights that the mean EAD and the mean Racine severity measurement, for each of the two.

asynchronous modes (PAs and ApAs) was substantially reduced in comparison to the two synchronous modes (PS and ApS), even taking into account standard error of measurement (SEM), indicated by the error bars. It is apparent that the greatest reduction may be achieved in the ApAs mode, although a large portion of the reduction can be achieved through use of asynchronicity.

The findings indicate that, when a plurality of electrodes are available for stimulation, adverse effects of electrical stimulation, such as epileptiform

afterdischarges, can be avoided by ensuring that electrodes are not stimulated synchronously for extended periods of time. Given that all of the stimulation strategies had substantially equivalent mean frequencies and pulse widths and therefore equal charge, it is shown that adjustment of the synchronicity between electrodes can override the proconvulsant nature of a stimulus. Following from this, by avoiding synchronous stimulation, higher current intensity electrical stimulation may be delivered than would have otherwise been possible.

Example 2: Experimental Methods and Materials

Female Wistar rats (250-400g) displaying spontaneous primarily generalised seizures were used as a second model in which therapeutic stimulation strategies were tested. The pre-surgical, anaesthetic and post-surgical routines used were similar to the routines of Example 1 , with the primary exception being the electrodes used were not microwire arrays. In this application, pairs of cortical screw electrodes or a monopolar wire-cortical screw pairing were implanted bilaterally to be aligned with Ml/SIHL in PC

22 the anteroposterior axis. The electrodes were placed approximately 1.5mm anterior and 3mm posterior to the Bregma.

All data were acquired at a sampling frequency of 3051Hz, Stimuli were typically delivered with a latency of 0.5 s to 2s from seizure onset with control seizures left unstimulated in a pseudo-random fashion. All delivered stimuli were between 0.2s and 0.5s in duration. Current intensities varied between 50μΑ and 650μΑ, depending on the electrode geometry used. Determination of EAD duration was performed offline using purpose written MATLAB analysis programs. EAD duration times were also independently verified by an experienced observer of clinical EEG.

.

Example 2: Results: Stimulation Strategy

Figure 9 shows the duration and treatment latency for the delivered stimuli and the total EAD duration for all trials with and without aperiodic/asynchronous, 500Hz (average rate approximately 370 Hz), ΙΟΟΟμβ stimuli. Each horizontal line represents a single trial with the position of the bar in the right hand chart indicating when each stimulus was delivered in relation to the seizure start. The left hand axes contain the non-stimulated trials EAD data and the axes of the right those from all stimulated trials. Figure 10 shows the same comparison but in this case the stimuli were set to lOOOHz (average rate approximately 740 Hz), 500μ5, aperiodic/asynchronous. In both figures the dashed lines show the percentage of trials in which the EAD has stopped by the 10 second mark. For the non-stimulated trials in both figures, between 50-60% of all seizures had ended by 10 seconds. For the 500Hz stimulated trials, this proportion had increased to approximately 85% of all seizures and for the l OOOHz stimulations greater than 90% of all seizures were of 10 seconds duration or less. Together with . the statistical significance presented below, these results clearly indicate a marked reduction in seizure duration when stimulation of this nature was, applied.

For the 500Hz stimulated trials, consistent with the fact that the stimuli were aperiodic/asynchronous, and the pulse width was 1000μ5 per phase, it should be noted that 500Hz was the maximum frequency of the electrical stimuli. The average frequency across the entire stimuli in this instance was approximately 370 Hz. Similarly, for the 1000Hz stimulated trials, consistent with the fact that the stimuli were aperiodic/asynchronous, and the pulse width was 500μ8 per phase, 1000Hz was also the maximum frequency of the electrical stimuli. The average frequency across the entire stimuli in this instance was approximately 740 Hz. Example 2: Statistical Significance of Seizure Abatement

The Mann- Whitney Rank Sum Test was again used to evaluate any statistical significance between the EAD populations of non-stimulated versus stimulated trials. For both 500Hz and 1000Hz stimulation significance (PO.001) was shown. Example 3: Experimental Methods and Materials; Results and Stimulation Strategy

The testing approach taken with respect to Example 2 was repeated, but, with reference to the plots of Figures 12 to 15, testing was performed by delivery of both aperiodic/asynchronous (Ap/As) and periodic/synchronous (P/S) stimuli, at both 125Hz and 500Hz, and with pulse width of 1000 μβεο, the stimulation current being maintained at 300μΑ. For the purpose of comparison, with reference to the plot of Figure 11 , testing was also performed upon delivery of periodic/synchronous stimuli, at 125Hz, with a pulse width of 300 μ$εο, and with the stimulation current maintained at 300 μΑ.

For the 500Hz stimulated trials, where the stimuli were aperiodic/asynchronous, since the width was ΙΟΟΟμε, it should be noted that 500Hz was the maximum frequency of the electrical stimuli. The average frequency across the entire stimuli in this instance was approximately 370 Hz.

The mean seizure duration and median seizure duration were calculated for all of the seizures shown in each plot, for non stimulated (NS) trials and stimulated (S) trials, and the details are set forth in Table 1. . I

24

Table 1

The results indicate a marked reduction in the duration of seizures upon delivery of stimulus_, with increased pulse width, and even greater reduction upon delivery of aperiodic/asynchronous stimulus with increased pulse width.

Example 3: Statistical Significance of Seizure Abatement

The Mann- Whitney Rank Sum Test was again used to evaluate any statistical significance between the EAD populations of non-stimulated versus stimulated trials. For all but the 125Hz, 300 8εο, P/S, 300 μΑ stimulation, significance (PO.OOl) was shown.

Example 4 Experimental Methods and Materials; Results and Stimulation Strategy

The testing approach taken with respect to Examples 2 and 3 was repeated, but testing was performed to analyse in yet further detail the respective effects, on seizure duration, of different stimulation paradigms, the different stimulation paradigms having varying frequencies, pulse widths and stimulation modes (P/S or Ap/As in this example). The results are represented graphically in Fig. 16, which shows a box and whisker plot of seizure duration vs stimulation paradigm. In this plot, the bottom and top of each box represents the 25th and 75th percentile (the lower and upper quartiles, P

25 respectively), the horizontal line within each box represents the 50th percentile (the median), and the bottom and top ends of the whiskers the 10^th and 90^lh percentiles, respectively. Outliers are each indicated by a cross (x).

From the plot, it can be seen that significant differences occur in seizure termination efficacy based on the choice of stimulation paradigm. From the very left hand side is shown the results for the Control, where detected seizures were not stimulated. Next to that is the results for what might be considered a typical stimulation paradigm for existing devices (125Hz, 300us, P/S), and at the very right hand edge is the results for one example of a stimulation paradigm according to the present disclosure (500Hz, ΙΟΟΟμββϋ, Ap/As). In between these two results, are the results for all the other permutations of the parameters.

The results continue to indicate that seizure termination efficacy does exist through use of stimulation paradigms in relation to this model of epilepsy. There is statistically significant interaction between the frequency and the stimulation mode (stimulation type) and between the pulse width and stimulation mode. Individually, pulse width and stimulation mode can be seen as the significant determinants of seizure termination efficacy, with, generally, a greater pulse width and/or the use of Ap/As (instead of P/S) both providing statistically significant improvements in stimulation seizure efficacy. On the other hand, individually, frequency is not seen as a significant determining factor.

Example 4: Statistical Significance of Seizure Abatement

The Mann- Whitney Rank Sum Test was again used to evaluate any statistical significance between the EAD populations of non-stimulated versus stimulated trials. For all stimulation combinations that utilised a pulse width of 1000 μ$εο significance was shown (PO.01). Combinations that Utilised both 1000 μ εο pulse widths and aperiodic/asynchronous stimulation mode showed the greatest reduction in median seizure duration (2.7 s) versus controls (8.9 s) of all stimulation paradigms. It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific 1473

P

26 embodiments without departing from the spirit or scope of the invention as broadly described. For example different electrode configurations may be used, leading to a change in electrode surface area and impacting the allowable charge densities in therapeutic stimuli. In such embodiments a change in intensity may occur to alter the dosage, while retaining functionality of the stimulus. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMS:

1. A device for counteracting seizure events in a mammalian brain, the device comprising:

a plurality of electrodes for delivering an electrical stimulus to the brain; and a stimulus generator which is configured to apply pulsatile electrical stimulus to each of the electrodes such that pulses are applied across the electrodes

asynchronously.

2. The device of claim 1, wherein the stimulus generator is configured to apply the pulses aperiodically at each electrode.

3. The device of claim 1 , wherein the stimulus generator is configured to apply the pulses periodically at each electrode.

4. The device of claim 1, 2 or 3, wherein the stimulus generator is configured to apply the pulses in a variable order across the electrodes.

5. The device of claim 4 when dependent on claim 2, wherein the variable order is a randomised variable order.

6. The device of claim 5, wherein the variable order is varied according to independent Poisson processes.

7. The device of claim 2, wherein the stimulus generator is configured to apply pulses having an inter-stimulus interval (ISI) that is consistent at any one of the electrodes and that is different across different electrodes.

8. The device of any one of claims 1 to 6, wherein the stimulus generator is configured to vary the ISI at one or more of the electrodes randomly.

9. The device of any one of the claims 1 to 6, wherein the stimulus generator is configured to vary the ISI at one or more of the electrodes according to independent Poisson processes.

10. The device of claim 9, wherein the ISI is varied at each electrode by independent Poisson processes having the same average pulse rate.

1 1. The device of claim 9, wherein the ISI is varied at each electrode by independent Poisson processes having different average pulse rates.

12. The device of any one of the preceding claims, wherein each electrode comprises a bipolar electrode pair.

13. The device of any one of the preceding claims comprising at least four of the electrodes.

14. The device of any one of the preceding claims comprising at least 5 of the electrodes.

15. The device of any one of the preceding claims comprising at least 10 of the electrodes.

16. The device of any one of the preceding claims wherein the pulses have a pulse width of greater than 300 μεβα

17. The device of claim 16, wherein the pulses have a pulse width of at least 500 sec.

18. The device of claim 16, wherein the pulses have a pulse width of at least 750 μ8βα

19. The device of any one of claims 16 to 18, wherein the pulses have a pulse width no greater than 1500 μεεο.

20. The device of any one of the preceding claims further comprising a monitor for monitoring electrical activity of the brain to detect a seizure event.

21. The device of claim 20, wherein the stimulus generator is configured to apply the electrical stimulus in response to a seizure event being detected by the monitor.

22. The device of any one of claims 1 to 20, wherein the stimulus generator is configured to apply the electrical stimulus substantially continuously or at regular intervals.

23. The device of any one of the preceding claims wherein the stimulus generator is configured to apply the electrical stimulus for at least 300ms.

24. The device of any one of claims 1 to 21 wherein the stimulus generator is configured to apply the electrical stimulus for between 300ms and 3000ms.

25. The device of any one of the preceding claims, wherein the stimulus generator is configured to apply pulses to the electrodes in cycles having cycle periods, such that, for at least 50% of the cycles, there is no gap in the cycle period of greater than 50% of the cycle period where no pulses are applied to at least one of the electrodes by the stimulus generator.

26. A method for counteracting seizure events in a mammalian brain, the method comprising applying electrical stimulus to the brain via a plurality of electrodes, the electrical stimulus comprising a train of electrical pulses, the application of the pulses across the electrodes being asynchronous.

27. The method of claim 26, wherein the pulses are applied aperiodically at each electrode.

28. The method of claim 26, wherein the pulses are applied periodically at each electrode.

29. The method of claim 26, 27 or 28, wherein the pulses are applied in a variable order across the electrodes.

30. The method of claim 29 when dependent on claim 27, wherein the variable order is a randomised variable order.

31. The method of claim 30, wherein the variable order is varied according to independent Poisson processes.

32. The method of claim 28, wherein the pulses have an inter-stimulus interval (ISI) that is consistent at any one of the electrodes and that is different across different electrodes.

33. The method of any one of claims 26 to 31 , wherein the ISI^' is varied at one or more of the electrodes randomly.

34. The method of any one of claims 26 to 31 , wherein the ISI is varied at one or more of the electrodes according to independent Poisson processes.

!

35. The method of claim 34 wherein the ISI is varied at each electrode by independent Poisson processes having the same average pulse rate.

36. The method of claim 34, wherein the ISI is varied at each electrode by independent Poisson processes having different average pulse rates.

37. The method of any one of claims 26 to 36, wherein the pulses have a pulse width of greater than 300 μ5ε£.

38. The method of any one of claims 26 to 37, wherein the pulses have a pulse width equal to or greater than 500 μεεα

39. The method of any one of claims 26 to 38, wherein the pulses have a pulse width equal to or greater than 750 μβεΰ.

40. The method of any one of claims 26 to 39, wherein the pulses have a pulse width equal to or less than 1500 μββα

41. The method of any one of claims 26 to 40, wherein pulses are applied across the electrodes in cycles having cycle periods, such that, for at least 50% of the cycles, there is no gap in the cycle period of greater than 50% of the cycle period where no pulses are applied to at least one of the electrodes.

42. A device for counteracting seizure events in a mammalian brain, the device comprising:

a plurality of electrodes for delivering an electrical stimulus to the brain; and a stimulus generator which is configured to apply pulsatile electrical stimulus to each of the electrodes, wherein the pulses have a pulse width of greater than 300 μ$εα

43. A method for counteracting seizure events in a mammalian brain, the method comprising applying electrical stimulus to the brain via a plurality of electrodes, the electrical stimulus comprising a train of electrical pulses, the pulses having a pulse width of greater than 300 μβεΰ.