Non-invasive prediction of an evoked potential

Abstract

Provided herein is a method for predicting: a proxy evoked potential pattern, EP pattern (124'), and / or one or more proxy evoked potential parameters, proxy EP parameters (126'), of the EP pattern (124'), and / or one or more parameters derivable from the one or more proxy EP parameters, proxy dEP parameters (126'), the method comprising: receiving test electroencephalogram data, test EEG data, of the test subject containing at least one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential, - determining from the at least one test FP EEG data segment, or one or more test FP EEG parameters determined from the FP EEG data segment, the proxy EP pattern, and / or the proxy EP parameter, and / or the proxy dEP parameter.

Description

[0001]Non-invasive prediction of an evoked potentialField of the invention The present invention is a field of neurological measurement, in particular of brain activity, inparticular of a prediction (measurement) of an evoked potential in a brain of a subject.Background to the invention In neurological disorders such as epilepsy, migraine, depression amongst others, the occurrence of the symptoms is related to the excitability of the brain. Metrics that reflect theexcitability of the brain are typically measured using evoked potentials. A stimulation pulse isprovided to the brain often using an implanted electrode, and the evoked potential responseof the neurons to this pulse is measured using the implanted electrode, which is related to theexcitability of the brain. However, the electrical stimulation pulse may elicit a seizure in thesubject, or there are limits as to a duration of measurement owing to the need for a pluralityof stimulation pulses over time. In addition, measurement and stimulation typically require animplanted electrode. It is an aim of the present invention to predict or measure evoked potentials in a subject withouta requirement for a stimulation pulse. It is the aim of the present invention to extract theinformation obtained through the use of evoked potential without the use of a stimulation pulse. Summary of the invention Described herein is a system comprising a processor configured for performing a method for predicting: -a proxy evoked potential pattern, EP pattern (124’), and / or- one or more proxy evoked potential parameters, proxy EP parameters (126’), of theEP pattern (124’); and / or -one or more parameters derivable from the one or more proxy EPP parameters,proxy dEP parameters (126’); of a test subject, wherein the proxy eP pattern (124’), the one or more proxy EP parameters (126’), or the one or more proxy dEP parameters (126’) is correlated to an EP pattern (124), one or more EP parameters (126), or one or more dEP parameters (126) respectively of the test subject; the method comprising:- receiving test electroencephalogram data, test EEG data, of the test subject containing atleast one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential;- determining from the- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG data segment;the proxy EP pattern, and / or the one or more proxy EP parameters, and / or the one or more proxy dEP parameters. Described herein is a method for measuring: -a proxy evoked field potential pattern, eFP pattern (124’), and / or- one or more proxy evoked field potential parameters, proxy eFP parameters (126’),of the eFP pattern (124’), and / or -one or more parameters derivable from the one or more proxy eFP parameters, proxydeFP parameters (126’), of a test subject, wherein the proxy eFP pattern (124’), the one or more proxy eFP parameters (126’), or the one or more proxy deFP parameters (126’) is correlated to an eFP pattern (124), one or more eFP parameters (126), or one or more deFP parameters (126) respectively as if measured for the test subject: -had the test subject received a stimulation pulse, and- had the test subject received a brain-implanted measurement probe for measuringthe eFP pattern, an eFP parameter or an deFP parameter respectively resulting from the stimulation pulse, the method comprising:- receiving test electroencephalogram data, test EEG data, of the test subject containing atleast one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked field potential,- determining from the- at least one test FP EEG data segment, or- one or more test FP EEG parameters determined from the FP EEG data segment,the proxy eFP pattern, and / or the proxy eFP parameter, and / or the proxy deFP parameter. The determining preferably comprises applying the at least one FP EEG data segments of the test subject, or the one or more test FP EEG parameters determined from the at least one FP EEG data segments to a solved or trained machine learning model.The solved or trained machine learning model is preferably a solved regression model or atrained neural network model.Where the solved or trained machine learning model is a solved regression model, the solvedregression model is preferably generated by:- adjusting variables of an unsolved regression model such that one or more record EP / dEP(eFP / deFP) parameters of a record data set approaches one or more record FP EEGparameters of the record data set, where the record data set comprises one or more record cycle datasets, each record cycle dataset comprising: -one or more record EP (eFP) parameters of the record subject and / or one or morerecord dEP (deFP) parameters derived from the record EP (eFP) parametersacquired during a record session, and- one or more record FP EEG parameters determined from the record FP EEG datasegment of the record subject acquired during the same record session.Where the solved or trained machine learning model is a trained neural network model, thetrained machine learning model is preferably generated by:- adjusting an untrained neural network model such that- one or more record EP / dEP ( eFP / deFP) parameters of a record data setapproaches one or more record FP EEG parameters of the record data set, or -one or more record EP / dEP (eFP / deFP) parameters of a record data setapproaches one or more record FP EEG data segments of the record data set, or -one or more record EP (eFP) patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or- one or more record EP (eFP) patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set, wherein the record data set comprises a plurality of record cycle datasets, each record cycle dataset comprising: -a record EP (eFP) pattern of the record subject acquired during a record sessionacquired during a record session, or -one or more record EP / dEP (eFP / deFP) parameters determined from a record EP(eFP) pattern of the record subject acquired during a record session acquired duringa record session, and -a record FP EEG data segments of the record subject acquired during thesame record session, or -one or more record FP EEG parameters determined from a record FP EEGdata segment(s) of the record subject acquired during the same record session.The one or more proxy or record EP / dEP (eFP / deFP) parameters preferably comprises oneor more of (A) peak slope (fEPSP_slope), (B) Population spike ((PS) amplitude of population spike), (C) peak amplitude (fEPSP), (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope.The one or more proxy or record EP / dEP (eFP / deFP) parameters preferably comprises (A)peak slope (fEPSP_slope), (B) Population spike (PS), and (C) peak amplitude (fEPSP), and optionally one or more of: (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope. The one or more test or record FP EEG parameters preferably comprises one or more of (a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) Phase-amplitude correlation. The method is an offline method. Further provided is a system comprising a processor configured for performing the method as described herein. Further provided is a computer program or computer program product having instructions which when executed by a computing device or system cause the computing device or system to perform the method as described herein. Further provided is a computer readable medium having stored thereon a computer program (product) having instructions which when executed by a computing device or system cause the computing device or system to perform the method as described herein.Further provided is a method for predicting:- a proxy evoked potential pattern, EP pattern (124’), and / or- one or more proxy evoked potential parameters, proxy EP parameters (126’), of theEP pattern (124’); and / or -one or more parameters derivable from the one or more proxy EPP parameters,proxy dEP parameters (126’); of a test subject, wherein the proxy eP pattern (124’), the one or more proxy EP parameters (126’), or the one or more proxy dEP parameters (126’) is correlated to an EP pattern (124), one or more EP parameters (126), or one or more dEP parameters (126) respectively of the test subject; the method comprising: -receiving test electroencephalogram data, test EEG data, of the test subjectcontaining at least one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential; -determining from the- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG datasegment; the proxy EP pattern, and / or the one or more proxy EP parameters, and / or the one or more proxy dEP parameters. The method according may include one of more of the limitations described herein. Themethod may be an offline method. The method may be a computer-implemented method.Figure LegendsFIG. 1 is a schematic illustration of the disclosure herein.FIG. 2 is a schematic illustration of the disclosure herein.FIG. 3 is a schematic illustration of an EP pattern with various (proxy or record or measured)EP parameters indicated.FIG. 4 panels A and B showing performance of a linear regression model for predicting peakamplitude (fEPSP), in respect of training data and test data respectively.FIG. 5 panels A and B showing performance of a linear for predicting Peak slope(fEPSP_slope), in respect of training data and test data respectively. FIG. 6 panels A and B showing performance of a linear regression model for predicting Population spike (PS), in respect of training data and test data respectively. FIG. 7 panels A and B showing performance of a linear regression model for predicting PS / fEPSP_slope (Neuronal excitability Index (NEI)) in respect of training data and test data respectively. Detailed description of invention Before the present system and method of the invention are described, it is to be understood that this invention is not limited to particular systems and methods or combinations described, since such systems and methods and combinations may, of course, vary. It is also to be understood that the terminology used herein is not intended to be limiting, since the scope of the present invention will be limited only by the appended claims. As used herein, the singular forms "a", "an", and "the" include both singular and plural referents unless the context clearly dictates otherwise. The terms "comprising", "comprises" and "comprised of" as used herein are synonymous ith "including", "includes" or "containing", "contains", and are inclusive or open-ended and do not exclude additional, non-recited members, elements or method steps. It will be appreciated that the terms "comprising", "comprises" and "comprised of" as used herein comprise the terms "consisting of", "consists" and "consists of". The recitation of numerical ranges by endpoints includes all numbers and fractions subsumed within the respective ranges, as well as the recited endpoints. The term "about" or “approximately” as used herein when referring to a measurable value such as a parameter, an amount, a temporal duration, and the like, is meant to encompass variations of + / -10% or less, preferably + / -5% or less, more preferably + / -1% or less, and still more preferably + / -0.1% or less of and from the specified value, insofar such variations are appropriate to perform in the disclosed invention. It is to be understood that the value to which the modifier "about" or “approximately” refers is itself also specifically, and preferably, disclosed. Whereas the terms “one or more” or “at least one”, such as one or more or at least one member(s) of a group of members, is clear per se, by means of further exemplification, the term encompasses inter alia a reference to any one of said members, or to any two or more of said members, such as, e.g., any ≥3, ≥4, ≥5, ≥6 or ≥7 etc. of said members, and up to all said members. All references cited in the present specification are hereby incorporated by reference in their entirety. In particular, the teachings of all references herein specifically referred to are incorporated by reference. Unless otherwise defined, all terms used in disclosing the invention, including technical and scientific terms, have the meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. By means of further guidance, term definitions are included to better appreciate the teaching of the present invention. In the following passages, different aspects of the invention are defined in more detail. Each aspect so defined may be combined with any other aspect or aspects unless clearly indicated to the contrary. In particular, any feature indicated as being preferred or advantageous may be combined with any other feature or features indicated as being preferred or advantageous. Reference throughout this specification to “one embodiment” or “an embodiment” means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to a person skilled in the art from this disclosure, in one or more embodiments. Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those in the art. For example, in the appended claims, any of the claimed embodiments can be used in any combination. In the present description of the invention, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration only of specific embodiments in which the invention may be practiced. Parenthesized or emboldened reference numerals affixed to respective elements merely exemplify the elements by way of example, with which it is not intended to limit the respective elements. Unless otherwise indicated, all figures and drawings in this document are not to scale and are chosen for the purpose of illustrating different embodiments of the invention. In particular the dimensions of the various components are depicted in illustrative terms only, and no relationship betweenthe dimensions of the various components should be inferred from the drawings, unless soindicated. It is to be understood that other embodiments may be utilised and structural or logical changes may be made without departing from the scope of the present invention. The following detailed description, therefore, is not to be taken in a limiting sense, and the scope of the present invention is defined by the appended claims.The present invention is based on a finding that a segment(s) of EEG data (FP EEG segment(128), FIG. 1) recorded after an evoked potential (EP) (124) has subsided, wherein the EPwas evoked in a record subject using a stimulation pulse (122) such as electrical, optical, auralstimulation and was recorded by a brain-implanted probe in the record subject, is indicative ofa pattern of the evoked potential (EP) (EP pattern (124)) and / or of one or more parameters(record EP parameters (126)) of the EP pattern in the record subject. The inventors havefound that FP EEG segments (128’) (FIG. 2) recorded in a test subject not exposed to anystimulation pulse, or in a test subject exposed to a stimulation pulse where the effects thereofhave faded, are indicative of or can be transformed (100) into a proxy EP pattern (124’) and / oroptionally into one or more proxy EP parameters (126’) for the test subject. The proxy EPpattern (124’) is correlated to a record EP pattern (124) as if measured using an implantedprobe in the test subject. The one or more proxy EP parameters (126’) are correlated to recordEP parameters (126) as if measured using an implanted probe in the test subject; accordingly,a proxy EP pattern (124’) and / or one or more proxy EP parameters can be measured in a testsubject without a need for an implanted probe and / or without a need for a stimulation pulse.Provided herein is a method for predicting:- a proxy evoked potential pattern, EP pattern (124’), and / or- one or more proxy evoked potential parameters, proxy EP parameters (126’), of theEP pattern (124’); and / or -one or more parameters derivable from the one or more proxy EPP parameters,proxy dEP parameters (126’); of a test subject, wherein the proxy eP pattern (124’), the one or more proxy EP parameters (126’), or the one or more proxy dEP parameters (126’) is correlated to an EP pattern (124), one or more EP parameters (126), or one or more dEP parameters (126) respectively of the test subject; the method comprising: -receiving test electroencephalogram data, test EEG data, of the test subjectcontaining at least one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential; -determining from the- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG datasegment; the proxy EP pattern, and / or the one or more proxy EP parameters, and / or the one or more proxy dEP parameters. In particular, provided herein is a method for predicting (or measuring): -a proxy evoked potential pattern, EP pattern, and / or- one or more proxy evoked potential parameters, proxy EP parameters, of the EPpattern, and / or- one or more parameters derivable from the one or more proxy EP parameters, proxydEP parameters,of a test subject, wherein the proxy EP pattern, the proxy EP parameter, the proxy dEPparameter is correlated to an EP pattern, an EP parameter or an dEP parameter respectively,as if measured for the test subject:- had the test subject received a stimulation pulse, and- had the test subject received a brain-implanted measurement probe for measuringthe EP pattern, an EP parameter or an dEP parameter respectively resulting from thestimulation pulse, the method comprising:- receiving test electroencephalogram data, test EEG data, of the test subject containing atleast one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evokedpotential, and- determining from the:- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG data segment;the proxy EP pattern, and / or the proxy EP parameter, and / or the proxy dEP parameter.Provided here is a system or processor configured for carrying out a method described herein.Normally measurement of EP pattern and / or of EP / dEP parameters require stimulation andusually an implanted electrode to measure and evoke the EP. With the present disclosure, thestimulation is no longer necessary, allowing extraction of the information from EP patternswithout the side-effects introduced by the stimulation as well as reducing the amount of implanted electrodes.Normally measurement of EP patterns and / or of EP / dEP parameters require visits to specialistcenters. With the present disclosure, data can be collected at local practices or at home.The present disclosure allows continuous or near-continuous monitoring. It does not need astimulation pulse, hence, monitoring can be performed near-continuously (e.g. every 0.1 seconds, every 10 seconds, 12 hours per day or longer) as there are no drawbacks to continuous recording contrary to the prior art which requires repeated stimulation pulses which do have drawbacks and are thus harder to use continuously. Continuous monitoring allows monitoring of brain dysfunctions whose level of severity isknown to be linked to one or more of the EP / dEP parameters, such as the Neuronal excitabilityIndex (NEI) e.g. epilepsy, tinnitus, pain, depression, migraine, mood disorders.Continuous monitoring allows prediction / warning of an event such as an epileptic event e.g.by monitoring a change in one or more of the EP / dEP parameters (e.g. rise in NEI), prior tothe event occurring. The patient can perform a self-intervention or retire to a place of safety. Alternatively or in addition, an intervention may be triggered automatically using a neuromodulator, e.g. for vagal nerve stimulation (VNS), electrical or optical stimulation. Some epileptic patients have a neurostimulator implanted; this means that record data (recordEP pattern and record EEG data segments) may be collected per patient fitted with an EEGsystem, for solving or training a machine learning model (a patient specific solved or trainedmachine learning model, see elsewhere herein) specific to that patient.The record data may be collected per patent from different populations of patients. Onepopulation of patients may be healthy, another population of patients may suffer from a dysfunction such as epilepsy. The population-specific record data may be used for solving ortraining a machine learning model (a population patient specific solved or trained machinelearning model, see elsewhere herein) for a population. Naive members of the populationgroup can hence benefit from continuous monitoring of proxy EP pattern or proxy EP / dEPparameters without needing an implanted stimulation electrode. In particular trends in (proxyEP pattern or proxy EP / dEP parameters over time are indicative of an increase or decrease in severity and / or a prediction of an event. A naive member of a population is one for whomno record data (no record FP pattern, no record EP / dEP parameters, no record EEG datasegments) has been collected.An evoked potential (EP), or sometimes called an evoked field potential (eFP), is amilliseconds long fragment of field potential measured in the brain of a subject in response to (evoked by) an external stimulus pulse. It is typically represented as an amplitude (voltage) / time chart. It is measured using a measurement probe (electrode). The external stimulus pulse may be an electrical or magnetic or optical pulse delivered to the brain using an probe (electrode) implanted into the brain of the subject or a surface probe, a sensory stimulus (auditory, tactile, visual / optical) or cognitive event. An evoked potential is well known in the art.The amplitude (voltage) / time chart of an evoked potential is known herein as an EP pattern.An exemplary proxy (124’) or record (124) EP pattern is shown in FIG. 3.The exemplary (proxy or record or measured) EP pattern has a baseline (BL) prior to thestimulation pulse (STP), a first peak (P1) having a positive potential, that leads over time intoa second peak (P2) having a negative potential, that leads over time into a third peak (P3) apositive potential. It is appreciated that a (proxy or record or measured) EP pattern maycontain one or more additional features, dependent on the test or record subject and / or on the external stimulus.Where there are multiple EP patterns (e.g. from multiple recording cycles mentionedelsewhere herein) of the subject in a session, they may be statistically combined (e.g.averaged) across the session to reduce noise. The statistical combination is preferably anaverage of time-consecutive EP patterns.The EP pattern may be a record or proxy EP pattern.A record EP pattern is an EP pattern that has been measured in a record subject for a purposeof solving or training a machine learning model as described elsewhere herein. It is measuredusing a measurement probe (electrode) implanted the brain of the subject.A record EP pattern may be determined from a statistical combination (e.g. average) ofmultiple record FP EEG segments patterns (e.g. from multiple recording cycles mentionedelsewhere herein) of the record subject in a record session. By averaging, noise is reduced.The statistical combination is preferably an average of time-consecutive record FP EEG segments.A proxy EP pattern is a pattern that is predictive of that EP pattern as if it had been measuredusing the measurement probe (electrode) implanted the brain of a test subject. A proxy EPpattern is correlated to the EP pattern as if it has been measured using a measurement probeimplanted the brain of the test subject. In the invention, the proxy EP pattern is determined inthe test subject from a segment of the EEG data (test FP EEG data segment) in which any evoked potential has ended or faded out, or where no evoked potential was evoked by a stimulation pulse.A proxy EP pattern of the subject parameter may be determined from a statistical combination(e.g. average) of multiple EEG FP segments patterns (e.g. from multiple recording cycles mentioned elsewhere herein) of the record subject in a record session. Where there aremultiple proxy EP patterns of the subject, they may be statistically combined (e.g. averaged)across the session, which reduces noise.An EP parameter is a descriptive of a feature of the EP pattern. Each EP parameter has ascalar value or level. Examples of EP parameters are discussed later in detail below andinclude (A) peak slope (fEPSP_slope), (B) Population spike (PS: amplitude of population spike), (C) peak amplitude (fEPSP), (D) peak latency, (E) power of high frequency (ripple)oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area underpopulation spike. Each EP parameter has a scalar value or level.A dEP parameter is a parameter derived from one or more of the EP parameters. By derived,it is meant the EP parameter is transformed, for example, by combination with another EPparameter. An example of a dEP parameter is [(B) PS / (A) fEPSP_slope] and is an indicatorof neuronal excitability index (NEI). Each dEP parameter has a scalar value or level.Neuronal Excitability Index (NEI) is a metric (scalar value) derived from the one or more EPparameters as a reflection of the neuronal excitability. A single parameter (EP parameter)such as (C) peak amplitude (fEP), (A) peak slope (fEP_slope), (H) area under population spike, (E) power of high frequency (ripple) oscillations is typically indicative of the NEI.Alternatively or in addition, a combination of EP parameters (i.e. a dEP parameter) such astwo or more of such as (A) peak slope (fEP_slope), (B) population spike (PS), (C) peak amplitude (fEP), (H) area under population spike, (E) power of high frequency (ripple) oscillations is typically indicative of the NEI. The combination selected depends on stimulation modality, stimulation location, and electrode location, as understood by the person skilled inthe art. Most indictive of NEI is the dEP parameter [(B) PS / (A) fEPSP_slope].Because an EP parameter, its derived dEP parameter, or both may be used or determined,the expression EP / dEP parameter is used herein, which refers to an EP parameter, a dEPparameter, or both.The EP / dEP parameter may be a record or proxy EP / dEP parameter.A record EP / dEP parameter is an EP / dEP parameter determined from the EP pattern that hasbeen measured in a record subject for a purpose of solving or training a machine learningmodel as described elsewhere herein. It is measured using a measurement probe (electrode)implanted the brain of the subject.A record EP / dEP parameter (e.g. (A) peak slope (fEPSP_slope)) may be determined from astatistical combination (e.g. average) of multiple EP patterns (e.g. from multiple recordingcycles mentioned elsewhere herein) of the record subject in a record session. The statisticalcombination is preferably an average of time-consecutive record EP patterns. Alternatively,multiple record EP / dEP parameters of the subject (e.g. multiple measurements of (A)) patterns(e.g. from multiple recording cycles mentioned elsewhere herein) in a record session, maybe statistically combined (e.g. averaged) across the record session. The statisticalcombination is preferably an average of time-consecutive record EP / dEP parameter.A proxy EP / dEP parameter is a parameter that is predictive of that EP / dEP parameter as if ithad been measured using the measurement probe (electrode) implanted the brain of a testsubject. A proxy EP / dEP parameter is correlated to the EP / dEP parameter as if it has beenmeasured using a measurement probe implanted the brain of the test subject. In the invention,the proxy EP / dEP parameter is determined in the test subject from a segment of the EEG data(test FP EEG data segment) in which any evoked potential has ended or faded out, or where no evoked potential was evoked by a stimulation pulse.A proxy EP / dEP parameter may be determined from a statistical combination (e.g. average)of multiple test EEG FP segments patterns (e.g. from multiple recording cycles mentioned elsewhere herein) of the test subject in a test session. The statistical combination is preferablyan average of time-consecutive record EP patterns. A proxy EP / dEP parameter (e.g. (A) peakslope (fEPSP_slope)) may be determined from a statistical combination (e.g. average) ofmultiple proxy EP patterns (e.g. from multiple recording cycles mentioned elsewhere herein)of the test subject in a test session. Alternatively, multiple proxy EP / dEP parameters of thetest subject (e.g. multiple predictions (or measurements) of (A)) patterns (e.g. from multiplerecording cycles mentioned elsewhere herein) in a test session, may be statistically combined(e.g. averaged) across the test session.As mentioned elsewhere herein, a proxy EP pattern is correlated to the EP pattern as if it hasbeen measured using a measurement probe implanted the brain of the test subject, and aproxy EP / dEP parameter is correlated to the EP / dEP parameter as if it has been measuredusing a measurement probe implanted the brain of the test subject. By correlated to, it is meant that a status of the brain observable from an -EP pattern and / or EP / dEP parameteris reflected in a status of the brain observable by the proxy EP pattern and / or proxy EP / dEPparameter.For, example, by correlated to, it is meant that, a change over time in the brain observablefrom -a change over time in an EP pattern and / or EP / dEP parameteris reflected by change over time in the proxy EP pattern and / or proxy EP / dEP parameter.Where a status of the brain changes, caused, for instance by a dysfunction, or an evolution(improvement or worsening) of a dysfunction, the change in status of the brain observable bythe proxy EP pattern and / or proxy EP / dEP parameter is correlated by a change in status ofthe brain observable from an EP pattern and / or EP / dEP parameter.The correlation is agnostic of the mode of stimulation (e.g. an electrical or magnetic or opticalpulse) in obtaining the EP pattern and / or EP / dEP parameter, because the brain status isgoverned by the physiology of the brain, and the different modes of stimulation merely resultin different intensities of the response.An electrical or magnetic or optical pulse will each invoke a similar EP pattern and / or EP / dEPparameter, and a changes to the status of the brain will be reflected also in a change to theEP pattern and / or EP / dEP parameter whether obtained by electrical or magnetic or opticalpulse stimulation. The proxy EP pattern and / or proxy EP / dEP parameter can hence be areflection of (correlated to) the EP pattern and / or EP / dEP parameter.Where a proxy EP / dEP parameter (126’) is correlated to a EP / dEP parameter (126), it isunderstood that the correlation is like-for like, e.g. proxy EP / dEP parameter (A) peak slope(fEPSP_slope) is correlated to a EP / dEP parameter (A) peak slope (fEPSP_slope).The method provides a proxy EP pattern and / or proxy EP / dEP parameter because it is moreuseful for the practitioner. The practitioner is familiar with analysing an EP pattern and / orEP / dEP parameter, from which he can subsequently make a diagnosis or monitor an evolutionof a dysfunction. FP EEG data segment is a portion (in time) of the EEG data devoid of influence of an evoked potential. By devoid of influence of an evoked potential it is meant that if a probe (electrode)would be implanted into the brain of the subject for measurement of evoked potential, anyevoked potential would be at a level not detectable by the probe. In practice, FP EEG data segment devoid of influence of an evoked potential is where the subject has not been not exposed to a stimulation pulse, or where the subject has been exposed to a stimulation pulse and a delay period is applied before the FP data EEG segment is extracted (usually after around 3 to 10 seconds after the stimulation pulse). An EP EEG data segment may have a duration of greater than 5 second, preferably 5 to 15 seconds.The EEG data segment is a time-voltage signal that can be transformed using a spectrogram,a wavelet transform or a Hilbert transform, or any other representation or transformation ofthe spectrogram using time information and / or frequency information of the spectrogram. A FP EEG data segment may be acquired for the test subject for a purpose of predicting (ormeasuring) the proxy FP pattern, and / or for a purpose of predicting (or measuring) the one ormore proxy EP / dEP parameters of the test subject; such FP EEG data segment(s) may beknown as a test FP EEG data segment(s). A FP EEG data segment may be acquired for a record subject for a purpose of solving ortraining a machine learning model as described elsewhere herein; such FP EEG datasegment(s) may be known as a record FP EEG data segment(s). A FP EEG parameter is a descriptive of a feature of the FP EEG data segment. One or more FP EEG parameters may be determined from the FP EEG data segment. Examples of FP EEG parameters are discussed later in detail below and include: (a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitudecorrelation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) Phase-amplitude correlation. The list is non-exhaustive. The one or more proxy EP / dEP parametersare determined from the one or more FP EEG parameters of the at least one dEP EEG datasegment. A FP EEG parameter may be measured on a test subject for a purpose of predicting (ormeasuring) one or more proxy EP / dEP parameters of the test subject; such FP EEGparameter may be known as a test FP EEG parameter(s). A test FP EEG parameter(s) (e.g. (a) Power spectral density) may be determined from a statistical combination (e.g. average) of multiple test FP EEG parameter(s) (e.g. from multiplerecording cycles mentioned elsewhere herein) of the test subject in a test session; each testFP EEG parameter(s) is determined from one EEG FP segment of the test subject in a test session. The statistical combination is preferably an average of time-consecutive test FP EEG parameter(s).A FP EEG parameter may be measured on a record subject for a purpose of solving or traininga machine learning model as described elsewhere herein; such FP EEG parameter may beknown as a record FP EEG parameter(s). A record FP EEG parameter(s) (e.g. (a) Power spectral density) may be determined from a statistical combination (e.g. average) of multiple record FP EEG parameter(s) (e.g. from multiple recording cycles mentioned elsewhere herein) of the record subject in a record session; each record FP EEG parameter(s) is determined from one EEG FP segment of the record subject in a record session. The statistical combination is preferably an average of time-consecutive record FP EEG parameter(s). A session is period of continuous acquisition lasting for a period of time (e.g.1 hour). It is notfragmented over time. A session may be a test session or a record session.A session may be a test session for a purpose of predicting (or measuring) one or more proxyEP patterns and / or one or more proxy EP / dEP parameters of the test subject for whom theone or more proxy EP patterns and / or one or more proxy EP / dEP parameters are not yetknown. A test session is period of continuous acquisition, during which at least one, preferably multiple FP data EEG segments are collected. The test session may not include one or more stimulation pulses. The test session may include one or more stimulation pulses, however, the field potentials evoked by the one or more stimulation pulse have ended or faded out during the test FP data EEG segment. Preferably, the test session is devoid of a stimulation pulse. A session may be a record session for a purpose of solving or training a machine learningmodel as described elsewhere herein. The record session is period of continuous acquisition,during which at least one, preferably multiple record FP data EEG segments, EP patterns,and optionally multiple record EP / dEP parameters are collected from the record subject. Therecord session is described in further detail below. The subject is preferably a human subject, however, other animals may be included such as vertebrates (e.g. birds) and mammals (e.g. cats, dogs). The subject may be a test subject or a record subject. The test subject is the subject for whom the one or more proxy EP / dEP parameters are to be measured and at least one test FP EEG data segment(s) is measured.The record subject is the subject for whom the one or more record EP / dEP parameters andrecord FP EEG data segment(s) are to be measured for a purpose of solving or training amachine learning model as described elsewhere herein.An exemplary (proxy (124’) or record (124)) EP pattern is shown in FIG.3. The exemplary EPpattern has a baseline (BL) prior to the stimulation pulse (STP), a first peak (P1) having a positive potential, that leads into a second peak (P2) having a negative potential, that leads into a third peak (P3) a positive potential. The first peak (P1) has a first apex (AI), the secondpeak (P2) has a second apex (nadir) (AII), the third peak (P3) has a third apex (AIII). It isappreciated that a (proxy or record or measured) EP pattern may contain one or moreadditional features, dependent on the test or record subject and / or on the external stimulus.The (proxy or record or measured) EP / dEP parameters of the EP pattern preferably includeone or more of (A) peak slope (fEPSP_slope), (B) Population spike ((PS) amplitude ofpopulation spike), (C) peak amplitude (fEPSP), (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope.The (A) peak slope (also known as fEPSP_slope) ((A) in FIG. 3) is a gradient of a best fitstraight line to the rising slope of the first peak (P1).The (B) population spike (PS) (also known as amplitude of population spike) ((B) in FIG.3) isan amplitude between the second apex (AII) and a straight line (SLI-III) contacting the first apex(AI) and third apex (AIII). The (C) peak amplitude (also known as fEPSP) ((C) in FIG.3) is an amplitude between thebaseline axis (BLX) and third apex (AIII).The (D) peak latency ((D) in FIG. 3) is a time duration between the time of the stimulationpulse and the time of the first apex (AI). The (E) power of high frequency (ripple) oscillations. The power of the high frequency (ripple) oscillations may be extracted by applying a high pass filter at 200Hz and consequently computing the power spectral density over the duration of the EP. The (F) coast-line index is determined from with n the index of the sample in the recorded signal.The (G) area under evoked potential (both (G)s in FIG. 3) is a sum of area of the first peak(P1) and third peak (P3).The (H) area under population spike ((H) in FIG.3) is an area of the second peak (P2) that extends to a straight line (SLI-III) contacting the first apex (AI) and third apex (AIII).The dEP parameters (proxy or record or measured) may include:(I) [(B) PS / (A) fEPSP_slope]. The ratio reflects input-output relationship of neurons and is anindicator of NEI.Each (proxy or record or measured) EP / dEP parameter has a scalar value or level. Preferably,the (proxy or record or measured) EP / dEP parameters include at least (A) Peak slope, (B)population spike, and (C) Peak amplitude.One or more (test or record) FP EEG parameters are extracted from the FP EEG datasegment. The one or more (test or record) FP EEG parameters are preferably one of more of(a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) Phase-amplitude correlation. The list is non-exhaustive.A level of one or more (a) to (i) is indicative of a level of the one or more proxy EP / dEPparameters. Alternatively or in addition, a change in one or more of (a) to (i) is indicative of achange in the one or more proxy EP / dEP parameters. Description of symbols used in the equations -^^^^(^^): sample of recorded field potential signal at time point t;- ^^^^: Frequency band j consisting of low and high cut-off frequencies;- ^^^^^^(^^): sample of recorded field potential signal, bandpass filtered in frequency bandj, at time point t; -^^[]: Hilbert transform; result is a complex signal;^^^^^^^^^^(): Extract angle from a complex signal; -angle of sample of recorded field potential signal, bandpass filtered infrequency band j, at time point t; -^ ^^^^^^^^^^^: Mean of recorded field potential signal, bandpass filtered in frequency band j;- ^^^^^^ : Standard deviation of recorded field potential signal, bandpass filtered infrequency band j; -Amplitude of sample of recorded field potential signal, bandpass filtered infrequency band j, at time point t; -angle of extracted amplitude of sample of recorded field potential signal,bandpass filtered in frequency band j, at time point t.The (a) power spectral density is an operator known in the art performed usingscipy.signal.welch implementation in Python. The (b) Log10 of power spectral density is a Log to base 10 of the (a) power spectral density.The (c) synchrony is calculated using a phase from a signal which is obtained using the Hilberttransform on the bandpass filtered EEG signal. See, for instance: https: / / www.pnas.org / doi / epdf / 10.1073 / pnas.1513716112The (d) Phase correlation (f1 f2) is calculated using the expression:^^^^(^^) = ^^^^^^^^^^^^^^^^^^^^(^^)^^^^^^(^^) = ^^^^^^^^^^^^^^^^^^^^(^^)^^ See, for instance: https: / / www.frontiersin.org / articles / 10.3389 / fnins.2019.00573 / fullThe (e) Amplitude correlation (f1 f2) is obtained by calculating the pearson correlationcoefficient squared (R²) between the two bandpass filtered signals with sigma the standard deviation of the filtered signal. Source: https: / / www.sciencedirect.com / science / article / pii / S0010482518304037 (but this is abasic and commonly used operation)The (f) Phase locking value is calculated according to the following expressions^^^^(^^) = ^^^^^^^^^^^^^^^^^^^^(^^)^^^^^^(^^) = ^^^^^^^^^^^^^^^^(^^)^^^^′^^(^^) = ^^^^^^^^^^^^^^^^^^^^(^^)^^ See, for instance: https: / / www.frontiersin.org / journals / neuroscience / articles / 10.3389 / fnins.2019.00573 / full The (g) mean vector length is calculated according to the following expressions See, for instance: https: / / www.frontiersin.org / articles / 10.3389 / fnins.2019.00573 / full The (h) modulation index is calculated according to the following expressions ^^^^(^^) = ^^^^^^^^^^^^^^^^^^^^(^^)^^^^^^(^^) = ^^^^^^^^^^^^^^^^(^^)^^Then, the phase is binned into 20° intervals and the mean amplitude of is calculated per bin.This is denoted < ^^^^ >^^^(^) with k indicating bin number. Then, the entropy is calculated: See, for instance: https: / / www.pnas.org / doi / 10.1073 / pnas.0810524105#supplementary-materials The phase locking value, mean vector length and modulation index are examples of (i) phase- amplitude correlation Preferably FP EEG parameter (a) is measured at one or more, preferably all of the following frequency bands: Delta: 0.5-4Hz; Theta: 4-8Hz; Alpha: 8-14Hz; Beta: 14-30Hz; Gamma: >30Hz. Preferably FP EEG parameter (b) is measured at one or more, preferably all of the following frequency bands: Delta: 0.5-4Hz; Theta: 4-8Hz; Alpha: 8-14Hz; Beta: 14-30Hz; Gamma: >30Hz; 0.01-400Hz; 0.01-100Hz Preferably FP EEG parameter (c) is measured at one or more, preferably all of the following frequency bands: Delta: 0.5-4Hz; Theta: 4-8Hz; Alpha: 8-14Hz; Beta: 14-30Hz; Gamma: >30Hz; 30-100Hz; 1-50Hz; 50-100Hz; 100-200Hz; 200-400Hz.Preferably FP EEG parameter (d) Phase correlation is measured at one or more, preferablyall of the following frequency bands: 0.5-4Hz to 4-8Hz; 8-14Hz to 14-30Hz; 14-30Hz to 30- 50Hz; 30-50Hz to 50-100Hz; 50-100Hz to 100-200Hz; 100-200Hz to 200-400Hz. Preferably FP EEG parameter (e) Amplitude correlation is measured at one or more, preferably all of the following frequency bands: 0.5-4Hz to 4-8Hz; 8-14Hz to 14-30Hz; 14- 30Hz to 30-50Hz; 30-50Hz to 50-100Hz; 50-100Hz to 100-200Hz; 100-200Hz to 200-400Hz. Preferably FP EEG parameter (f) Phase locking value is measured at one or more, preferably all of the following frequency bands: 0.5-4Hz to 4-8Hz; 8-14Hz to 14-30Hz; 14-30Hz to 30- 50Hz; 30-50Hz to 50-100Hz; 50-100Hz to 100-200Hz; 100-200Hz to 200-400Hz. Preferably FP EEG parameter (g) Mean vector length is measured at one or more, preferably all of the following frequency bands: 0.5-4Hz to 4-8Hz; 8-14Hz to 14-30Hz; 14-30Hz to 30- 50Hz; 30-50Hz to 50-100Hz; 50-100Hz to 100-200Hz; 100-200Hz to 200-400Hz. Preferably FP EEG parameter (h) Modulation index is measured at one or more, preferably all of the following frequency bands: 0.5-4Hz to 4-8Hz; 8-14Hz to 14-30Hz; 14-30Hz to 30- 50Hz; 30-50Hz to 50-100Hz; 50-100Hz to 100-200Hz; 100-200Hz to 200-400Hz.Preferably, the one or more (proxy or record or measured) EP / dEP parameters comprisesone or more of:(A) peak slope (fEPSP_slope), (B) Population spike (PS), (I) PS / fEPSP_slope, and (C) peak amplitude (fEPSP), andoptionally one or more of: (D) peak latency, (E) power of high frequency (ripple) oscillations,(F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike.Preferably, the (proxy or record or measured) EP / dEP parameters comprises (A) peak slope(fEPSP_slope), (B) Population spike (PS), and optionally (C) peak amplitude (fEPSP),and optionally one or more of (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope.The (proxy or record or measured) EP / dEP parameters may comprise:(A) peak slope (fEPSP_slope), (B) Population spike (PS), and (C) peak amplitude (fEPSP), (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, and (I) PS / fEPSP_slope.Preferably, the one or more (proxy or record or measured) FP EEG parameters comprisesone or more of: (a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) phase-amplitude correlation.Preferably, (proxy or record or measured) FP EEG parameters comprises: (a) Powerspectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, and (h) Modulation index. The determining from the one or more test FP EEG data segments or the one or more test FP EEG parameters, -the proxy EP pattern and / or- the one or more proxy EP / dEP parameterspreferably comprises applying the one or more test FP EEG data segments or the one or moretest FP EEG parameters of the test subject to a solved or trained machine learning model.The solved or trained machine learning model may be a solved regression model, having asan input: -the one or more test FP EEG data segments of the test subject; or- the one or more test FP EEG parameters of the test subject;and as an output: -the proxy EP pattern of the test subject; or- the one or more proxy EP / dEP parameters of the test subject.The solved or trained machine learning model may be a trained neural network model, havingas an input: -the one or more test FP EEG data segments of the test subject; or- the one or more test FP EEG parameters of the test subject;and as an output: -the proxy EP pattern of the test subject; or- the one or more proxy EP / dEP parameters of the test subject.The solved or trained machine learning model is preferably generated (solved or trained) by:- adjusting (variables of) an unsolved or untrained trained machine learning model such that:- one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set,wherein the record data set comprises a plurality of record cycle datasets, each record cycledataset comprising: -a record EP pattern of the record subject acquired during a record session, or- one or more record EP / dEP parameters of the record subject acquired during arecord session, and -one or more record FP EEG parameters determined from the record FPEEG data segment(s) of the record subject acquired during the same recordsession, or -a record FP EEG data segments of the record subject acquired during thesame record session. Preferably, the determining comprises:- applying:- the one or more test FP EEG data segments of the test subject or- the one or more test FP EEG parameters of the test subject,to the solved or trained machine learning model, and- calculating from an output of the solved or trained machine learning model,- the proxy EP pattern of the test subject; or- the one or more proxy EP / dEP parameters of the test subject.The machine learning model may be a linear correlation, random forest regression, MLPregressor, a machine learning model including deep learning algorithms such as CNN’s,LSTM’s and / or GRU’s, or any type of model that is trained, refined, optimized or solved using record (training) data set from a record subject.Where the solved or trained machine learning model is a solved regression model, the solvedregression model is preferably generated by:- adjusting (variables of) the unsolved regression model such that one or more recordEP / dEP parameters of a record data set approaches one or more record FP EEGparameters of the record data set, where the record data set comprises one or more record cycle datasets, each record cycle dataset comprising: -one or more record EP / dEP parameters of the record subject acquired during arecord session, and- one or more record FP EEG parameters determined from the record FP EEG datasegment of the record subject acquired during the same record session. Preferably, the determining comprises:- applying the one or more test FP EEG parameters of the test subject to the solvedregression model, and- determining from an output of the regression model, the one or more proxy EP / dEPparameters of the test subject. Examples of suitable regression models and protocols for solving them are known in the art, for instance, from Cherian and Kanaga, Journal of Neuroscience Methods, Volume 369, 1March 2022, 109483 and Vattikonda et al, Communications Biology volume 4, Articlenumber: 1244 (2021). Non-limiting examples of regression models include Lasso regression (linear regression with L1 regularization), Linear SVM, and Bayesian regression. The solved regression model may be a test-patient specific solved regression model, a TPS solved regression model, wherein the TPS solved regression model has been solved by: -adjusting (variables of) a TPS unsolved regression model such that one or morerecord EP / dEP parameters a record data set approaches one or more record FPEEG parameters of a record data set, where the record data set comprises one or more record cycle datasets of a record subject, and the record subject is the test patient.The solved regression model may be a population-specific solved regression model, a PSsolved regression model, wherein the PS solved regression model has been solved by:- adjusting (variables of) an unsolved PS regression model such that one or morerecord EP / dEP parameters of a record data set approaches one or more FP EEGparameters of a record data set, where the record data set comprises multiple recordcycle datasets from a population of record subjects (each different from the test subject). The population of record subjects may be a population based on one or more of age, gender, ethnicity, dysfunction history.Where the solved or trained machine learning model is a trained neural network model, thetrained neural network model is preferably generated by:- adjusting (variables of) an untrained neural network model such that- one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or- one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set, wherein the record data set comprises a plurality of record cycle datasets, each record cycle dataset comprising: -a record EP pattern of the record subject acquired during a record session acquiredduring a record session, or -one or more record EP / dEP parameters determined from a record EP pattern of therecord subject acquired during a record session, and -one or more record FP EEG parameters determined from the record FPEEG data segment(s) of the record subject acquired during the same record session, or- a record FP EEG data segment of the record subject acquired during thesame record session. Preferably, the determining comprises:- applying- the one or more test FP EEG data segments of the test subject; or- the one or more test FP EEG parameters of the test subject;to the trained neural network model, and- calculating from an output of the trained neural network model:- the proxy EP pattern of the test subject; or- the one or more proxy EP / dEP parameters of the test subject.Examples of suitable neural network models and protocols for training them are known inthe art, for instance, from Cherian and Kanaga, Journal of Neuroscience Methods, Volume 369, 1 March 2022, 109483. Examples of neural network models include Artificial Neural Network (ANN) / Multi Layer Perceptron (MLP), Recurrent Neural Networks (RNN), Convolutional Neural Network (CNN), Gated Recurrent Units (GRU), Long Short Term Memory (LSTM). The trained neural network model may be a test-patient specific trained neural networkmodel, a TPS trained neural network model, wherein the TPS trained neural network modelhas been trained by adjusting (variables of) an untrained neural network algorithm such that:- one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set, and the record subject is the test patient.The trained neural network model may be a population-specific trained neural networkmodel, a PS trained neural network model, wherein the PS trained machine learning model has been trained by:- adjusting (variables of) a untrained neural network model such that- one or more record EP / deFP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record eFP / deFP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set,wherein the record data set comprises multiple record cycle datasets from a population ofrecord subjects (each different from the test subject). The population of record subjects may be a population based on one or more of age,gender, ethnicity, dysfunction history. Each of the one or more measured EP parametersmay be statistically combined across the population (e.g. by averaging), and / or each of theone or more FP EEG parameters may be statistically combined across the population (e.g. by averaging), and / or each of the one or more FP EEG data segments may be statistically combined across the population (e.g. by averaging).For collection of EEG data from the (test or record) subject during a session, the subject isdisposed with an EEG measurement module (e.g. headset) comprising a plurality of EEGelectrodes in order to record the electroencephalogram of the subject in order to determine the at least one FP EEG data segment and optionally the at least one FP EEG parameters therefrom. The EEG electrode may be a surface electrode (patch) or may be implanted (intracranial or cortical). The EEG electrode may be a scalp, subgaleal, epidural, subduralelectrode. Examples of commercially available EEG measurement modules include thosemanufactured by Brain products, Nicolet, Micromed.For acquisition of the record EP pattern and / or record EP / dEP of the record subject during arecord session, typically, the record subject has a measurement probe (electrode)implanted. The electrode is typically a depth electrode located in a cortical area, intracranialarea or scalp. In a preferred aspect, the electrode is implanted in the dentate gyrus in thehippocampus of the record subject. The stimulation may be any stimulation type from electrical to optical stimulation (with lightpulses), transcranial magnetic stimulation or sensory input. In a preferred aspect, thestimulation probe is located in a perforant path which projects from entorinal cortex to thedentate gyrus.The record session is a continuous recording session lasting for a period of time (e.g. 1 to 5hours). It is not fractionated over time e.g. not fractionated over days or weeks. A record(training) session is a session comprising multiple, preferably consecutive, recording cycles. In each recording cycle, the record subject receives a (preferably one) stimulation pulse, and -an EP is recorded of the record subject responsive to the stimulation pulse;- an EEG is recorded of the record subject, and an (preferably one) FP EEG datasegment and therefrom one or more FP EEG parameters are extracted from the recorded EEG. A duration of a recording cycle may be 15 to 25 seconds. From the record session of the record subject, the record cycle dataset is generated. A (one) record cycle dataset of the record subject comprises: -an EP pattern determined from multiple recording cycles of the record session;and / or -at least one record EP / dEP parameter (e.g. one or more of A) to I)) determined fromthe one EP pattern;and -a corresponding record FP EEG data segment determined from multiple recordingcycles of the record subject acquired during the same record session; and / or -one or more record FP EEG parameters (e.g. one or more of a) to h)) determinedfrom the corresponding record FP EEG data segment of the record subject acquired during the same record session. The present method / system may be used for treatment of a dysfunction. In such case, theproxy EP pattern and / or proxy EP / dEP parameters may be linked to the dysfunction. Forexample, where is dysfunction is known to be linked to a presence of one or more (record orproxy) proxy EP patterns and / or EP / dEP parameters that are outside certain healthy ranges,a subject may be treated by one or more interventions that restore the one or more proxy EPpattern and / or EP / dEP parameters to be within the certain healthy ranges for the subjectbeing treated. Thus a closed feedback loop system may be provided that continually monitors the proxy EPpattern and / or proxy EP / dEP parameters according to the methods described herein, anddynamically adjusts a level of treatment such that the one or more proxy EP patterns and / orproxy EP / dEP parameters is within the certain healthy range. By dynamically adjusts, it ismeant that the level (intensity) of treatment is adjusted at regular intervals responsive toprediction (or measurement) of the proxy EP parameters at regular intervals. The regularintervals may be, for instance, on a scale of seconds, minutes, hours or days, preferably minutes. An example of a treatment may include neuromodulation (implanted and / or external electrodes).The present method may be used to screen for (proxy or draft) EP / dEP parameters that arelinked to a dysfunction. The (online or offline) method comprises:- obtaining set of proxy EP / dEP parameters, a dysfunction parameter set, using a methoddescribed herein from a population groups of subjects suffering from the dysfunction,- obtaining set of proxy EP / dEP parameters, a healthy parameter set, using a methoddescribed herein from a population group of healthy subjects,- determining, by comparing the dysfunction parameter set with the healthy parameter set,the proxy EP / dEP parameters linked to the dysfunction, andoptionally an intensity range(s) of each proxy EP / dEP parameter linked to thedysfunction where the dysfunction is (likely to be) present, and / or optionally an intensity range(s) of each proxy EP / dEP parameter linked to thedysfunction where the dysfunction is (likely to be) absent. Examples of treatable and / or screenable dysfunctions include epilepsy, tinnitus, pain, migraine, depression, mood disorders, anxiety disorders, obsessive-compulsive, Tourette syndrome, obesitas, cardiac diseases.Further provided is a system comprising:- an EEG measurement module comprising one or more EEG electrodes,- a processing unit, configured to carry out a method according to any one of theprevious claims.The EEG measurement module is operatively connected to the processing unit.Further provided is a system comprising: -an EEG measurement module comprising one or more EEG electrodes,- a processing unit, configured to carry out a method according to any one of theprevious claims, -a neuromodulator system (comprising a controller and lead containing at least onestimulation electrode) for delivery of electrical or optical stimulation to a test subject for regulation of epilepsy, -wherein the processing unit is further configured to instruct the neuromodulatorsystem to deliver stimulation (electrical and / or optical) for regulation of epilepsyresponsive to a change in a level of one or more proxy EP / dEP parameters.The EEG measurement module is operatively connected to the processing unit. Further provided is a system comprising: -an EEG measurement module comprising one or more EEG electrodes,- a processing unit, configured to carry out a method according to any one of theprevious claims, and- a neuromodulator system for delivery of stimulation (electrical and / or optical) to a testsubject for regulation of epilepsy, wherein the processing unit is further configured to acquire from the subject one or morerecord cycle datasets of the subject, for solving or training the (patient specific) machinelearning model thereby allowing calculation of one or more proxy EP / dEP parameters of thetest subject. The EEG measurement module is operatively connected to the processing unit. The neuromodulator system for delivery of electrical stimulation typically comprises a controller and lead containing at least one stimulation electrode for contacting neural tissue. The neuromodulator system for delivery of optical stimulation typically comprises a controllerand a light output unit for delivery of optical stimulation. The optical stimulation is preferablydelivered using an implanted waveguide to locally illuminate optical tissue.A method for generating a solved or trained machine learning model for determining- a proxy EP pattern and / or- one or more proxy EP / dEP parametersof a test subject in a test session, the method comprising:- adjusting (variables of) an unsolved or untrained machine learning model such that:- one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or- one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set, wherein the record data set comprises a plurality of record cycle datasets, each record cycle dataset comprising: -a record EP pattern of the record subject acquired during a record session acquiredduring a record session, or -one or more record EP / dEP parameters of the record subject acquired during arecord session, and -one or more record FP EEG parameters determined from the record FPEEG data segment(s) of the record subject acquired during the same record session, or -a record FP EEG data segments of the record subject acquired during thesame record session. The present method may be performed online or offline. By online, it is meant that the one ormore proxy EP / dEP parameters are determined from the at least one FP EEG data segmentwhile the EEG data of the test subject is being received from the test subject. By offline, it ismeant that the one or more (proxy EP / dEP parameters are determined from the at least oneFP EEG data segment while the EEG data of the test subject is being received from a datastorage device (e.g. hard-drive, cloud storage).The present method does not require invasive steps for predicting (or measuring) the one ormore proxy EP parameters, and / or one or more proxy dEP parameters, of the test subject.The present method is performed using a computer. The present method is a computer- implemented method. Further provided is a computing device or system configured for performing the method described herein. Further provided is a computer program or computer program product having instructions which when executed by a computing device or system cause the computing device or system to perform (each of the steps of) a method as described herein Further provided is a computer readable medium having stored thereon a computer program (product) having instructions which when executed by a computing device or system cause the computing device or system to perform (each of the steps of) a method as described herein. A data stream which is representative of a computer program or computer program product having instructions which when executed by a computing device or system cause the computing device or system to perform (each of the steps of) the method as described herein.The computing device or system or processor is preferably a standard computer systemsuch as an Intel Architecture IA-32 based computer system 2, and implemented as programming instructions of one or more software modules stored on non-volatile (e.g., hard disk or solid-state drive) storage associated with the corresponding computer system. However, it will be apparent that at least some of the steps of any of the described processes could alternatively be implemented, either in part or in its entirety, as one or more dedicated hardware components, such as gate configuration data for one or more field programmable gate arrays (FPGAs), or as application-specific integrated circuits (ASICs), for example. The method or system may produce an output that is:- displayed on a screen, or- saved to a file.According to one aspect, a system or kit is provided for use in a method described herein. According to one aspect, a use of a system is provided for use in a method described herein, wherein said device comprises at least one EEG electrode. According to one aspect, a use of a system is provided for use in a method described herein, wherein said device comprises an EEG measurement module. The system or kit or system or kit for use is a method may further comprise one or more of:- a computer system for executing the method described herein;- an EEG measurement module.ExampleA test subject was exposed to a series of stimulation pulses using an implanted electrode,and a series of corresponding EP patterns were recorded using also the implantedelectrode. At the same time, EEG data was captured, and a series of corresponding EEG data segments were extracted. A linear regression model was generated using:Record EP parameters (A) Peak slope (fEPSP_slope), and (B) Population spike (PS), (C)Peak amplitude (fEPSP), and Record FP EEG parameters: (a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) Phase-amplitude correlation. Notations: -y: Output variable (in this case (d)EP parameter)- ^^^: Input variables (in this the FP parameters. There are N in total)- ^^^,^: Nth model parameter of iteration i- ^^^: Mean square prediction error in iteration i over samples j- ^^: Learning rateStart iteration i: Update all model parameters as follows: Repeat from “start iteration” until convergence of model parameters (i.e. the update becomes insignificant for all parameters)The linear regression model was then tested using measured record FP EEG parametersand measured record EP parameters measured from the record subject using an implantedprobe that had not been used in the model generation (naive parameters). The linearregression model EP parameters outputted predictions of Peak slope (fEPSP_slope),Population spike (PS), Peak amplitude (fEPSP) for the training dataset (measured), andthese were compared with the EP parameters of Peak slope (fEPSP_slope), (Populationspike (PS), and Peak amplitude (fEPSP) for test dataset (unknown EP parameters).The results showed a good correlation between predicted EP parameters (predicted label)and measured record EP parameters (actual label) for both training and test datasets. Asshown in FIG. 4 for Peak amplitude (fEPSP), the p-value was 0.000 (Panel A, training)compared with the p-value of 0.0015 (Panel B, test); the R2-value was 0.2261 (Panel A,training) compared with the R2-value of 0.0869 (Panel B, predicted). As shown in FIG. 5 forPeak slope (fEPSP_slope), the p-value was 0.000 (Panel A, training) compared with the p- value of 0.0000 (Panel B, test); the R2-value was 0.6348 (Panel A, training) compared withthe R2-value of 0.5759 (Panel B, test). As shown in FIG.6 for Population spike (PS), the p-value was 0.000 (Panel A, training) compared with the p-value of 0.0000 (Panel B, test) ; the R2-value was 0.5853 (Panel A, training) compared with the R2-value of 0.4814 (Panel B, test). As shown in FIG.7 for [PS / fEPSP_slope] (NEI), the p-value was 0.000 (Panel A, training) compared with the p-value of 0.0000 (Panel B, test) ; the R2-value was 0.5901 (Panel A, training) compared with the R2-value of 0.5901 (Panel B, test).

Claims

Claims1. A system comprising a processor configured for performing a method for predicting:- a proxy evoked potential pattern, EP pattern (124’), and / or- one or more proxy evoked potential parameters, proxy EP parameters (126’), of theEP pattern (124’); and / or- one or more parameters derivable from the one or more proxy EPP parameters,proxy dEP parameters (126’);of a test subject, wherein the proxy eP pattern (124’), the one or more proxy EP parameters(126’), or the one or more proxy dEP parameters (126’) is correlated to an EP pattern (124),one or more EP parameters (126), or one or more dEP parameters (126) respectively of thetest subject; the method comprising: -receiving test electroencephalogram data, test EEG data, of the test subjectcontaining at least one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential; -determining from the- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG datasegment; the proxy EP pattern, and / or the one or more proxy EP parameters, and / or the one ormore proxy dEP parameters.

2. The system according to claim 1, wherein the determining comprises applying the at leastone FP EEG data segment of the test subject, or the one or more test FP EEG parametersdetermined from the at least one FP EEG data segments to a solved or trained machinelearning model.

3. The system according to claim 2, wherein the solved or trained machine learning model isa solved regression model or a trained neural network model.

4. The system according to claim 3, wherein the solved or trained machine learning model isa solved regression model, and the solved regression model is generated by:- adjusting variables of an unsolved regression model such that one or more record EP / dEPparameters of a record data set approaches one or more record FP EEG parameters of the record data set, where the record data set comprises one or more record cycle datasets, each record cycle dataset comprising: -one or more record EP parameters of the record subject and / or one or more recorddEP parameters derived from the record EP parameters acquired during a recordsession; and- one or more record FP EEG parameters determined from the record FP EEG datasegment of the record subject acquired during the same record session.

5. The system according to claim 3, wherein the solved or trained machine learning model isa trained neural network model, the trained neural network model is generated by:- adjusting an untrained neural network model such that- one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG parameters of the record data set, or -one or more record EP / dEP parameters of a record data set approaches one ormore record FP EEG data segments of the record data set, or -one or more record EP patterns of a record data set approaches one or morerecord FP EEG parameters of the record data set, or- one or more record EP patterns of a record data set approaches one or morerecord FP EEG data segments of the record data set, wherein the record data set comprises a plurality of record cycle datasets, each record cycle dataset comprising: -a record EP pattern of the record subject acquired during a record session acquiredduring a record session, or -one or more record EP / dEP parameters determined from a record EP pattern of therecord subject acquired during a record session acquired during a record session, and -a record FP EEG data segments of the record subject acquired during thesame record session, or- one or more record FP EEG parameters determined from a record FP EEGdata segment(s) of the record subject acquired during the same record session.

6. The system according to any one of claims 1 to 5, wherein the one or more proxy or recordEP / dEP parameters comprises one or more of (A) peak slope (fEPSP_slope), (B) Populationspike ((PS) amplitude of population spike), (C) peak amplitude (fEPSP), (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope.

7. The system according to any one of claims 1 to 6, wherein the one or more proxy orrecord EP / dEP parameters comprises (A) peak slope (fEPSP_slope), (B) Population spike(PS), and (C) peak amplitude (fEPSP), and optionally one or more of: (D) peak latency, (E) power of high frequency (ripple) oscillations, (F) Coast-line index, (G) area under evoked Field Potential, (H) area under population spike, (I) PS / fEPSP_slope.

8. The system according to any one of claims 1 to 7 wherein the one or more test or recordFP EEG parameters comprises one or more of (a) Power spectral density, (b) Log10 of power spectral density, (c) Synchrony, (d) Phase correlation, (e) Amplitude correlation, (f) Phase locking value, (g) Mean vector length, (h) Modulation index, (i) Phase-amplitude correlation.

9. A computer program or computer program product having instructions which when executed by a computing device or system cause the computing device or system to perform a method as defined in any one of claims 1 to 8.

10. A computer readable medium having stored thereon a computer program (product) havinginstructions which when executed by a computing device or system cause the computingdevice or system to perform a method as defined in any one of claims 1 to 8.

11. A method for predicting:- a proxy evoked potential pattern, EP pattern (124’), and / or- one or more proxy evoked potential parameters, proxy EP parameters (126’), of theEP pattern (124’); and / or- one or more parameters derivable from the one or more proxy EPP parameters,proxy dEP parameters (126’); of a test subject, wherein the proxy eP pattern (124’), the one or more proxy EP parameters (126’), or the one or more proxy dEP parameters (126’) is correlated to an EP pattern (124), one or more EP parameters (126), or one or more dEP parameters (126) respectively of the test subject; the method comprising: -receiving test electroencephalogram data, test EEG data, of the test subjectcontaining at least one data segment, test FP EEG data segment, of the test EEG data, wherein the test FP EEG data segment is a portion of the test EEG data devoid of influence of an evoked potential; -determining from the- at least one test FP EEG data segment; or- one or more test FP EEG parameters determined from the FP EEG datasegment; the proxy EP pattern, and / or the one or more proxy EP parameters, and / or the one or more proxy dEP parameters.

12. The method according to claim 11, including the limitations of any one of claims 2 to 8.

13. The method according to claim 11 or 12 that is an offline method.

14. The method according to any one of claims 11 to 13 that is a computer-implementedmethod.

Images