Classifying instances of parkinson's disease

Abstract

A system (28) includes multiple electrodes (22) configured to record respective signals produced, by a brain of a patient (12) suffering from a particular subtype of Parkinson's Disease (PD), in response to magnetic stimulation of the brain. The system further includes a processor (32) configured to receive the signals, to identify a subset, of multiple subtypes of PD, to which the particular subtype belongs, by processing the signals, and to output an output indicating the subset of the subtypes. Other embodiments are also described.

Description

[0001] CLASSIFYING INSTANCES OF PARKINSON’S DISEASE

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] The present application claims priority to US Provisional Application 63 / 537,797, filed September 11, 2023, entitled “Classifying instances of Parkinson’s disease,” whose disclosure is incorporated herein by reference.

[0004] FIELD OF EMBODIMENTS OF THE INVENTION

[0005] The present invention relates to the field of medical diagnostics, particularly with respect to Parkinson’s Disease (PD).

[0006] BACKGROUND

[0007] The use of electrophysiological measurements to characterize and monitor brain network activity has been used extensively over the last seven decades. Electrophysiological measurements can be generally divided into two groups of parameters: network integrity, i.e., the connectivity and coherence of the network, and network plasticity, also referred to as “neuroplasticity” or “brain plasticity.” Network connectivity depends on the synchronous activation of neurons. Network coherence refers to the level of synchrony between two or more brain regions and is used to determine the strength of connectivity between specific brain regions. Neuroplasticity is an ability of the brain to continuously adapt its functional and structural organization to changing requirements. Neuronal plasticity allows the brain to reorganize neuronal networks in response to environmental stimulation, to remember information and to recover from brain and spinal cord injuries. Neuronal plasticity is essential to the establishment and maintenance of brain circuitry.

[0008] Magnetic simulation, e.g., transcranial magnetic stimulation (TMS), is a non-invasive brain stimulation method that allows the study of human cortical function in vivo. Using magnetic stimulation for examining human cortical functionality is enhanced by combining such stimulation with registration of an electrical evoked response, such as an electroencephalograph (EEG). EEG provides an opportunity to directly measure the cerebral response to TMS, measuring the cortical TMS evoked potential (TEP). An important feature of TEP topography is that even though only one cortical hemisphere is stimulated, bi-hemispheric EEG responses are evoked with different features. TMS -evoked activity propagates from the stimulation site ipsilaterally via association fibers, contralaterally via transcallosal fibers, and to subcortical structures via projection fibers. A single TMS pulse delivered over the primary motor cortex (Ml) results in a sequence of positive and negative EEG peaks at specific latencies (typically, negative peaks at 45ms (N45) and 100ms (N100) after stimulation, and positive peaks at 60ms (P60) and 180ms (Pl 80) after stimulation). This pattern of response indicates synaptic activity. These TMS-evoked cortical potentials last for up to 300ms both in the vicinity of the stimulation and in remote interconnected brain areas.

[0009] PD is a progressive degenerative disorder of the central nervous system. Symptoms of PD include tremor, rigidity, reduced mobility, and cognitive impairment.

[0010] PD is a heterogeneous disorder with variable clinical characteristics and imaging findings, as well as different genetic and other underlying neuropathologic mechanisms. As a result, various classification (or “subtyping”) schemes for PD have been proposed. For example, Lewis, S. J. G., et al., "Heterogeneity of Parkinson’s disease in the early clinical stages using a data driven approach," Journal of Neurology, Neurosurgery & Psychiatry 76.3 (2005): 343-348, whose disclosure is incorporated herein by reference, defines four subtypes of PD, which are referred to herein as earlier disease onset (EDO), tremor dominant (TD), non-tremor dominant (NTD), and rapid disease progression (RDP). In the TD subtype, tremor is the dominant symptom. In NTD, on the other hand, tremor is mild or nonexistent, whereas, typically, cognitive impairment is significant. In RDP, the disease progresses quickly, e.g., such that the patient dies within 10 years of the first symptoms, although cognitive impairment is typically absent.

[0011] Maidan, Inbal, et al., "A multimodal approach using TMS and EEG reveals neurophysiological changes in Parkinson's disease," Parkinsonism & Related Disorders 89 (2021): 28-33, whose disclosure is incorporated herein by reference, provides evidence to various alterations in neurophysiological measures manifested in TEPs of patients with PD.

[0012] SUMMARY

[0013] In any given instance of PD, it may be important to identify the subtype of PD to which the instance belongs, or at least to rule out one or more PD subtypes, so as to determine the best course of treatment for the patient. However, conventional techniques for doing this have limited accuracy and / or practicality.

[0014] To address this challenge, embodiments of the present invention provide a system including multiple electrodes and a processor. The electrodes are configured to record respective signals produced, by a brain of a patient suffering from a particular subtype of PD, in response to magnetic stimulation of the brain including, typically, magnetic stimulation of an occipital cortex of the brain. The processor is configured to receive the signals and, by processing the signals, to identify a subset, of multiple subtypes of PD, to which the particular subtype belongs. The processor is further configured to output an output indicating the subset of the subtypes.

[0015] For example, the processor may be configured to ascertain whether the particular subtype is the rapid disease progression (RDP) subtype rather than, for example, the non-tremor dominant (NTD) subtype. The output may thus indicate whether the particular subtype is the RDP subtype.

[0016] Typically, the processor is configured to process the signals by computing one or more measures of neurophysiological activity exhibited in the signals and inputting the measures of neurophysiological activity to a model, such as a logistic regression model. The model is configured to output the subset of PD subtypes and, optionally, an estimated PD duration and / or PD stage of progression for the patient, which also help determine the best course of treatment.

[0017] There is therefore provided, in accordance with some embodiments of the present invention, a system, including multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain, and a processor. The processor is configured to receive the signals and, by processing the signals, identify a subset, of multiple subtypes of PD, to which the particular subtype belongs. The processor is further configured to output an output indicating the subset of the subtypes.

[0018] In some embodiments, the processor is configured to identify the subset by ascertaining whether the particular subtype is a rapid disease progression (RDP) subtype, and the output indicates whether the particular subtype is the RDP subtype.

[0019] In some embodiments, the processor is configured to identify the subset by ascertaining whether the particular subtype is a non-tremor dominant (NTD) subtype, and the output indicates whether the particular subtype is the NTD subtype.

[0020] In some embodiments, at least some of the signals were produced in response to magnetic stimulation of an occipital cortex of the brain.

[0021] In some embodiments, the processor is further configured to compute, by processing the signals, an estimated PD duration for the patient, and the output indicates the estimated duration.

[0022] In some embodiments, the processor is further configured to compute, by processing the signals, an estimated PD stage of progression for the patient, and the output indicates the estimated stage of progression.

[0023] In some embodiments, the processor is configured to process the signals by: computing one or more measures of neurophysiological activity exhibited in the signals, and inputting the measures of neurophysiological activity to a model.

[0024] In some embodiments, the model includes a logistic regression model.

[0025] In some embodiments, the measures of neurophysiological activity include a latency measure, and the processor is configured to compute the latency measure by: computing at least one average signal, by averaging at least some of the signals, identifying a timing of a greatest local maximum of the average signal within a predefined time window, and computing the latency measure based on the timing.

[0026] In some embodiments, the measures of neurophysiological activity include an interhemispheric connectivity measure, which quantifies a similarity between a right-side response of a right side of the brain to the magnetic stimulation and a left- side response of a left side of the brain to the magnetic stimulation.

[0027] There is further provided, in accordance with some embodiments of the present invention, a method, including receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The method further includes, by processing the signals, identifying, by the processor, a subset, of multiple subtypes of PD, to which the particular subtype belongs, and outputting an output indicating the subset of the subtypes.

[0028] There is further provided, in accordance with some embodiments of the present invention, a computer software product including a tangible non-transitory computer-readable medium in which program instructions are stored. The instructions, when read by a processor, cause the processor to receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The instructions further cause the processor to identify, by processing the signals, a subset, of multiple subtypes of PD, to which the particular subtype belongs, and to output an output indicating the subset of the subtypes.

[0029] There is further provided, in accordance with some embodiments of the present invention, a system including multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain, and a processor. The processor is configured to receive the signals, to compute an estimated PD duration for the patient by processing the signals, and to output an output indicating the estimated PD duration.

[0030] There is further provided, in accordance with some embodiments of the present invention, a method including receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The method further includes, by processing the signals, computing an estimated PD duration for the patient, and outputting an output indicating the estimated PD duration.

[0031] There is further provided, in accordance with some embodiments of the present invention, a computer software product including a tangible non-transitory computer-readable medium in which program instructions are stored. The instructions, when read by a processor, cause the processor to receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The instructions further cause the processor to compute an estimated PD duration for the patient by processing the signals, and to output an output indicating the estimated PD duration.

[0032] There is further provided, in accordance with some embodiments of the present invention, a system including multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain, and a processor. The processor is configured to receive the signals, to compute an estimated PD stage of progression for the patient by processing the signals, and to output an output indicating the estimated PD stage of progression.

[0033] There is further provided, in accordance with some embodiments of the present invention, a method including receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The method further includes, by processing the signals, computing an estimated PD stage of progression for the patient, and outputting an output indicating the estimated PD stage of progression.

[0034] There is further provided, in accordance with some embodiments of the present invention, a computer software product including a tangible non-transitory computer-readable medium in which program instructions are stored. The instructions, when read by a processor, cause the processor to receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain. The instructions further cause the processor to compute an estimated PD stage of progression for the patient by processing the signals, and to output an output indicating the estimated PD stage of progression.

[0035] The present invention will be more fully understood from the following detailed description of embodiments thereof, taken together with the drawings, in which:

[0036] BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Fig. 1 is a schematic illustration of a clinician performing a diagnostic procedure on a patient using a diagnostic system, in accordance with some embodiments of the present invention;

[0038] Fig. 2 is a flow diagram for a method for identifying a subtype of PD from which a patient suffers, or at least one subtype of PD from which the patient does not suffer, in accordance with some embodiments of the present invention;

[0039] Fig. 3 is a flow diagram for a measure-computing step, in accordance with some embodiments of the present invention; and

[0040] Figs. 4, 5A, and 5B show experimental results demonstrating the efficacy of diagnosing a PD subtype in accordance with some embodiments of the present invention. DETAILED DESCRIPTION

[0041] OVERVIEW

[0042] In any given instance of PD, it may be important to identify the subtype of PD to which the instance belongs, or at least to rule out one or more PD subtypes, so as to determine the best course of treatment for the patient. However, conventional techniques for doing this have limited accuracy and / or practicality.

[0043] To address this challenge, embodiments of the present invention analyze the neurophysiological response of the patient to magnetic stimulation, which, as the present inventors realized, may vary between PD subtypes. In particular, embodiments of the present invention provide a diagnostic system 28, shown in Fig. 1. The system comprises a magnetic stimulation device 20, multiple electrodes 22, and a computer processor 32. The magnetic stimulation device is configured to stimulate one or more cortexes of the patient’s brain. Following each stimulation, the electrodes record the resulting evoked-potential signals produced by the brain in response to the stimulation. The processor processes the signals so as to identify the PD subtype from which the patient suffers, or at least to identify one or more PD subtypes from which the patient does not suffer.

[0044] For example, per one method, which is shown in Fig. 2 with reference number 38, the processor computes multiple measures of neurophysiological activity exhibited in the signals, and then inputs these measures to a model configured to discriminate between multiple PD subtypes based on the measures. The computed measures may include, for example, a latency measure, which is related to the timing of the greatest local maximum of the response to the stimulation within a predefined time window. As demonstrated in the experimental results shown in Fig. 4, the present inventors have found that this measure is discriminating with respect to the rapid disease progression (RDP) subtype of PD, particularly with stimulation of the occipital cortex. In particular, the RDP subtype is associated with a lower latency (i.e., an earlier greatest local maximum), relative to other PD subtypes.

[0045] In some embodiments, by processing the signals, the processor also estimates how long the patient has been afflicted with symptoms of PD (referred to herein as “PD duration” or “disease duration”) and / or the extent to which these symptoms have progressed (referred to herein as “PD stage of progression” or “disease stage of progression”). For example, the aforementioned model may be configured to perform one or both of these estimations based on the input measures of neurophysiological activity. The present inventors have found that an interhemispheric connectivity measure, which quantifies the similarity between the patient’s right-side response and the patient’s left-side response, is particularly useful for estimating disease stage of progression, particularly with stimulation of the (right and left) occipital cortex.

[0046] SYSTEM DESCRIPTION

[0047] Reference is initially made to Fig. 1, which is a schematic illustration of a clinician 10 performing a diagnostic procedure on a patient 12 using a diagnostic system 28, in accordance with some embodiments of the present invention.

[0048] System 28 comprises a magnetic stimulation device 20 configured for placement near the head of patient 12, e.g., by virtue of being placed over a headpiece 30 worn over the head. Device 20 comprises one or more coils configured to generate a magnetic field, which stimulates activity within the brain of patient 12. By virtue of being configured to stimulate the brain from outside the body of the patient, device 20 may be referred to as a transcranial magnetic stimulation (TMS) device, though it is noted that device 20 may differ from conventional TMS devices (which are usually designated for treatment, rather than diagnosis).

[0049] In general, device 20 may be placed over any suitable portion of the patient’s brain, such as a frontal cortex, occipital cortex, parietal cortex, temporal cortex, or motor cortex of the brain.

[0050] System 28 further comprises multiple electrodes 22 configured to record the signals produced by the brain in response to the stimulation. Electrodes 22 may be coupled to the patient’s head via a low-impedance adhesive material. Alternatively, electrodes 22 may be coupled to headpiece 30 such that, when headpiece 30 is fittingly placed over the patient’ s head, the electrodes contact the head. Collectively, electrodes 22 may be referred to as an electroencephalograph (EEG) detector.

[0051] System 28 further comprises a control unit 24 comprising a computer processor 32, a signal generator 34, and other circuitry. The other circuitry typically comprises analog-to- digital (A / D) conversion circuitry and, optionally, denoising circuitry. Typically, control unit 24 is connected to stimulation device 20 via a cable 36, and to electrodes 22 via one or more other wired connections that, for ease of illustration, are not shown in Fig. 1. Alternatively, computer processor 32 may belong to an external device, such as a laptop, in communication with control unit 24, e.g., via a universal serial bus (USB) cable. Processor 32 is configured to drive signal generator 34 to generate electrical signals, which flow through device 20 via cable 36 (or via wireless transfer). The flow of the electrical signals through device 20 causes the device to generate a magnetic field, which in turn evokes signals (or “potentials”) in the patient’s brain. These signals are recorded by electrodes 22 and passed from the electrodes, via respective leads (not shown) or via wireless transfer, to the control unit. After optional denoising and digitization within control unit 24, processor 32 receives the signals.

[0052] Processor 32 is further configured to process the signals, as described in detail hereinbelow. In response to processing the signals, the processor outputs an output. For example, the processor may display the output on a display 26.

[0053] In some cases, prior to the diagnostic procedure, patient 12 is known to suffer from Parkinson’s Disease (PD), but the particular subtype of PD from which the patient suffers is unknown. In such cases, the output from processor 32 may specify a (strict) subset of a larger set of subtypes to which the particular subtype belongs. (The subset may consist of a single subtype, i.e., the output may specify the particular subtype from which the patient suffers.) Alternatively, the output may specify a subset of one or more subtypes to which the particular subtype does not belong, thereby implicitly specifying a subset of one or more subtypes to which the particular subtype belongs. (It is noted that the relevant subtype(s) need not be explicitly specified in the output. For example, each subtype may be specified by a corresponding alphanumeric code.)

[0054] In some embodiments, the set of subtypes includes the four subtypes described above in the Background: earlier disease onset (EDO), tremor dominant (TD), non-tremor dominant (NTD), and rapid disease progression (RDP). Alternatively, one of the subtypes is the union of two or more of the above. For example, the set of subtypes may include NTD and non-NTD, the latter subtype being the union of EDO, TD, and RDP. As another example, the set of subtypes may include RDP and non-RDP, the latter subtype being the union of EDO, TD, and NTD.

[0055] For example, it may not be known whether the patient suffers from the tremor dominant (TD), non-tremor dominant (NTD), or rapid disease progression (RDP) subtype of PD. Alternatively, it may be known (based on the absence of tremor) that the patient does not suffer from the TD subtype, but not known whether the patient suffers from the NTD subtype or the RDP subtype. (Typically, there is no doubt as to whether the patient suffers from the earlier disease onset (EDO) subtype, as this may be readily ascertained from the age of the patient.) In such a case, based on a processing of the patient’s evoked potentials, the processor may ascertain whether the patient has the RDP subtype or whether the patient has the NTD subtype. In response thereto, the processor may output an output indicating whether the patient has the RDP subtype or whether the patient has the NTD subtype.

[0056] Alternatively, the subtypes may be defined in accordance with any other suitable scheme. For example, the set may include two subtypes: PD with cognitive impairment, and PD without cognitive impairment.

[0057] In other cases, it is not known, prior to the diagnostic procedure, whether patient 12 suffers from PD at all. In such cases, based on the patient’s evoked potentials, the processor may ascertain whether the patient has PD, e.g., as described in the reference to Maidan et al. cited above in the Background. For example, in some embodiments, the processor differentiates between PD and other conditions that cause parkinsonism. If the patient has PD, the processor may further identify a subset of subtypes to which the patient’ s subtype belongs, based on either the same evoked potentials or other evoked potentials.

[0058] In some embodiments, when performing a diagnosis, the processor computes respective likelihoods for the relevant options. The processor then checks if the greatest likelihood exceeds a predefined threshold, or exceeds the next-greatest likelihood by a predefined threshold, and if so, chooses the corresponding option. For example, the processor may compute a first likelihood that the patient has a given PD subtype, and a second likelihood that the patient does not have the given PD subtype. Provided the first likelihood exceeds a predefined threshold, the processor may determine that the patient has the given PD subtype.

[0059] In some such embodiments, the processor’s output includes one or more of the computed likelihoods, or one or more confidence measures derived from such likelihoods. For example, when outputting an indication that the patient has a given PD subtype, the processor may specify the likelihood of this diagnosis being correct and, optionally, the likelihood(s) for one or more alternative diagnoses.

[0060] In view of the foregoing, it is noted that any reference herein, including in the claims, to the processor “identifying” that the patient’s PD subtype belongs to a given subset of subtypes, does not indicate that the identification is made by the processor with 100% certainty.

[0061] In general, processor 32 may be embodied as a single processor, or as a cooperatively networked or clustered set of processors. For example, processor 32 may be embodied as a local processor, which is disposed within control unit 24, cooperatively networked (e.g., over the Internet) with a remote processor. The local processor may stimulate the patient’s brain (by driving signal generator 34) and receive the resulting evoked signals. The local processor may communicate the signals to the remote processor, and the remote processor may process the signals as described in detail hereinbelow.

[0062] The functionality of processor 32 may be implemented solely in hardware, e.g., using one or more fixed-function or general-purpose integrated circuits, Application-Specific Integrated Circuits (ASICs), and / or Field-Programmable Gate Arrays (FPGAs). Alternatively, this functionality may be implemented at least partly in software. For example, processor 32 may be embodied as a programmed processor comprising, for example, a central processing unit (CPU) and / or a Graphics Processing Unit (GPU). Program code, including software programs, and / or data may be loaded for execution and processing by the CPU and / or GPU. The program code and / or data may be downloaded to the processor in electronic form, over a network, for example. Alternatively or additionally, the program code and / or data may be provided and / or stored on non-transitory tangible media, such as magnetic, optical, or electronic memory. Such program code and / or data, when provided to the processor, produce a machine or special-purpose computer, configured to perform the tasks described herein.

[0063] PROCESSING THE EVOKED POTENTIALS

[0064] Reference is now made to Fig. 2, which is a flow diagram for a method 38 for identifying a subtype of PD from which patient 12 (Fig. 1) suffers, or at least one subtype of PD from which the patient does not suffer, in accordance with some embodiments of the present invention. Typically, method 38 is performed in cases in which it is known that the patient has PD, but the particular subtype of the patient’s PD is unknown.

[0065] Method 38 begins at a signal-receiving step 40, at which processor 32 receives respective signals recorded by electrodes 22 (Fig. 1). As described above with reference to Fig. 1, these signals were produced, by the brain of the patient, in response to magnetic stimulation of the brain.

[0066] Subsequently to receiving the signals, the processor checks, at a checking step 42, whether more stimulations are to be performed. For example, the processor may perform checking step 42 by processing an input from the clinician. If more stimulations are to be performed, the processor returns to signal-receiving step 40, and receives the evoked potentials from the next stimulation.

[0067] In general, any cortex of the patient’s brain, such as the right or left frontal, parietal, temporal, motor, or occipital cortex, may be stimulated any number of times. In particular, the present inventors have found that for diagnosing the presence or absence of the RDP or NTD subtype, stimulating an occipital cortex may be particularly helpful. Hence, the clinician may perform several (e.g., 8-12) stimulations of the right occipital cortex and several (e.g., 8-12) stimulations of the left occipital cortex, such that the processor receives multiple (e.g., 16-24) signals from each electrode.

[0068] Upon ascertaining that no more stimulations are to be performed, the processor, by processing the signals, identifies a (strict) subset, of multiple subtypes of PD, to which the patient’s PD subtype belongs. As noted above, this subset may consist of a single subtype, i.e., the processor may identify the patient’s subtype. For example, the processor may ascertain whether the particular subtype is the RDP subtype, i.e., the processor may ascertain that the patient has the RDP subtype, or, alternatively, a non-RDP subtype. Alternatively, for example, the processor may ascertain whether the particular subtype is the NTD subtype, i.e., the processor may ascertain that the patient has the NTD subtype, or, alternatively, a non-NTD subtype.

[0069] Typically, in processing the signals, the processor first performs a measure-computing step 43, at which the processor computes one or more measures of neurophysiological activity exhibited in the signals. Subsequently, at an inputting step 50, the processor inputs the measures of neurophysiological activity to a model, such as a logistic regression model or neural network, configured to identify the subset of PD subtypes based on the measures.

[0070] Finally, at an outputting step 52, the processor outputs an output indicating the subset of PD subtypes to which the patient’s subtype belongs.

[0071] For example, the output may indicate whether the patient’s subtype is the RDP subtype. In other words, the output may specify that the patient suffers from the RDP subtype, or (equivalently) that the patient does not suffer from the TD subtype or the NTD subtype. Alternatively, the output may specify that the patient does not suffer from the RDP subtype, or (equivalently) that the patient suffers from either the TD subtype or the NTD subtype.

[0072] As another example, the output may indicate whether the patient’s subtype is the NTD subtype. In other words, the output may specify that the patient suffers from the NTD subtype, or (equivalently) that the patient does not suffer from the TD subtype or the RDP subtype. Alternatively, the output may specify that the patient does not suffer from the NTD subtype, or (equivalently) that the patient suffers from either the TD subtype or the RDP subtype. In some embodiments, alternatively or additionally to identifying a subset of PD subtypes, the processor, by processing the signals, computes an estimated PD duration for the patient, i.e., an estimated amount of time for which the patient has had PD. Alternatively or additionally, by processing the signals, the processor may compute an estimated PD stage of progression for the patient, i.e., an estimated degree to which the patient's PD symptoms have progressed. For example, the aforementioned model may be configured to compute the estimated duration and / or stage of progression based on the measures input at inputting step 50.

[0073] In such embodiments, the output at outputting step 52 may indicate the estimated duration and / or stage of progression.

[0074] Similarly to the Hoehn and Yahr (H&Y) scale, the estimated disease stage of progression may be computed as a number between 1 and 5, where 1 corresponds to minimal progression of symptoms and 5 corresponds to maximal progression. Alternatively, the estimated disease stage of progression may be computed with respect to any other suitable numerical or qualitative scale.

[0075] Reference is now made to Fig. 3, which is a flow diagram for measure-computing step

[0076] 43, in accordance with some embodiments of the present invention.

[0077] In some embodiments, measure-computing step 43 begins with a signal-averaging step

[0078] 44. At signal- averaging step 44, the processor computes an average signal for each of the stimulated cortexes, by averaging the received signals over the electrodes and stimulations. (Typically, the processor first computes an average for each electrode by averaging over the stimulations, and then averages these per-electrode averages.) For example, assuming repeated stimulations of both the right and left occipital cortex were performed, the processor may compute an average signal for the right occipital cortex and another average signal for the left occipital cortex.

[0079] Next, in some embodiments, at a connectivity-measure-computing step 45, the processor computes an interhemispheric connectivity measure, which quantifies the similarity between the response of the right side of the patient’s brain to the stimulation and the response of the left side of the patient’s brain to the stimulation. For example, assuming that, at signalaveraging step 44, the processor calculated average signals for the right and left occipital cortexes, the processor may compute the interhemispheric connectivity measure by comparing these two average signals to one another. In the event that average signals were computed for multiple cortex types - e.g., right and left occipital, and also right and left frontal, parietal, or motor - the processor computes the interhemispheric connectivity measure separately for each cortex type.

[0080] The present inventors have found that, at least for occipital-cortex stimulation, the interhemispheric connectivity measure may be particularly helpful for estimating the patient’s disease stage of progression. For example, the present inventors have found that, at least for occipital-cortex stimulation, the interhemispheric connectivity measure is correlated with the H&Y scale.

[0081] In some embodiments, alternatively or additionally to connectivity-measure-computing step 45, the processor, at an other-measure-computing step 46, computes one or more other measures of neurophysiological activity from each of the average signals. Subsequently, at a measure-averaging step 48, the processor computes an average of each of these measures. (In the event that average signals were computed for multiple cortex types, the processor computes a separate average for each cortex type.)

[0082] In some embodiments, the other measures of neurophysiological activity include a latency measure. To compute this measure, the processor first identifies the greatest local maximum of the average signal within a predefined time window. Next, the processor computes the latency measure based on the time, relative to the start of the stimulation or the start of the time window, at which the greatest local maximum occurs. For example, the processor may set the latency measure equal to this time.

[0083] In some embodiments, the predefined time window begins 10-80 ms (e.g., 30, 40, 50, or 60 ms), 100-160 ms (e.g., 120 or 140 ms), or 150-190 ms (e.g., 170 ms) after the start of the stimulation. Alternatively or additionally, the predefined time window may have a duration of 20-100 ms, such as 40, 60, or 80 ms. Thus, for example, the predefined time window may span any of the following intervals: 30-90 ms, 40-100 ms, 60-120 ms, 120-200 ms, 140-220 ms, 170- 210 ms.

[0084] Thus, for example, the processor may compute a respective latency measure for each of the right and left occipital cortexes, and then average these latency measures at measureaveraging step 48. As shown in Fig. 4, which is described below, the present inventors have found that, at least for occipital-cortex stimulation, the latency measure may help differentiating the RDP subtype from the TD and NTD subtypes. In particular, the RDP subtype may exhibit a lower latency measure, relative to the other two subtypes.

[0085] Alternatively or additionally to a latency measure, the other measures of neurophysiological activity may include a cortical excitability measure. This measure may be computed based on the area under the average signal or a statistic of the amplitude of the signal, such as the mean average deviation of the amplitude. For example, the measure may equal the area or the statistic of the amplitude. Thus, for example, the processor may compute a respective cortical excitability measure for each of the right and left occipital cortexes, and then average these cortical excitability measures at measure-averaging step 48.

[0086] Alternatively or additionally, the other measures of neurophysiological activity may include a wide waveform adherence measure. This measure may be computed based on the similarity between the average signal and a benchmark signal. Thus, for example, the processor may compute a respective wide waveform adherence measure for each of the right and left occipital cortexes, and then average these wide waveform adherence measures at measureaveraging step 48.

[0087] Alternatively to averaging each of the measures at measure-averaging step 48, the processor may compute a weighted average, a maximum, a minimum, or any other statistic for each of the measures.

[0088] Subsequently to performing measure-computing step 43, the processor inputs the computed interhemispheric connectivity measure(s) and / or the computed measure average(s) to the model at inputting step 50, as described above with reference to Fig. 2.

[0089] In general, the model may be computed using logistic regression or any other suitable technique, based on measures of neurophysiological activity exhibited in the evoked potentials of a suitable group of PD patients. The “ground truth” PD subtype and / or disease duration and / or disease stage of progression for each these patients may be identified, for example, using the Movement Disorder Society-Sponsored Revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS) assessment, a subjective evaluation of disease duration and / or severity of symptoms, and / or a medical evaluation.

[0090] EXPERIMENTAL DATA

[0091] Reference is now made to Fig. 4, which shows experimental results demonstrating the efficacy of diagnosing a PD subtype in accordance with some embodiments of the present invention.

[0092] Fig. 4 shows mean and standard-error bars for a latency measure computed for RDP, TD, and NTD patients, following magnetic stimulation of the right and left occipital cortex. As can be seen from the time values along the horizontal axis, the latency measure is a "late-phase" latency measure, in that the predefined time window is relatively late.

[0093] As shown in Fig. 4, following stimulation of the right and left occipital cortex, the late- phase latency for TD or NTD is significantly greater than the late-phase latency for RDP. Thus, these experimental results support evaluating whether the patient has RDP by inputting the patient’s late-phase latency, following occipital stimulation, to a model.

[0094] Reference is now made to Figs. 5A-B, which show additional experimental results demonstrating the efficacy of diagnosing a PD subtype in accordance with some embodiments of the present invention.

[0095] Fig. 5A shows a receiver operating characteristic curve for a logistic regression model configured to discriminate between the NTD subtype and the non-NTD subtypes based on the wide waveform adherence, interhemispheric connectivity, cortical excitability, and late-phase latency measures, following occipital stimulation. Fig. 5B shows a receiver operating characteristic curve for a logistic regression model configured to discriminate between the RDP subtype and the non-RDP subtypes based on the wide waveform adherence, interhemispheric connectivity, cortical excitability, and late-phase latency measures, following occipital stimulation. In view of the relatively high specificity and sensitivity attained by the models, it can be concluded that wide waveform adherence, interhemispheric connectivity, cortical excitability, and late-phase latency following occipital stimulation are effective measures for discriminating between the NTD subtype and the non-NTD subtypes and between the RDP subtype and the non-RDP subtypes.

[0096] Embodiments of the invention described herein can take the form of a computer program product accessible from a computer-usable or computer-readable medium (e.g., a non- transitory computer-readable medium) providing program code for use by or in connection with a computer or any instruction execution system, such as computer processor 32 (Fig. 1). For the purpose of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation medium. Typically, the computer-usable or computer readable medium is a non-transitory computer-usable or computer readable medium. Examples of a computer-readable medium include a semiconductor or solid-state memory, magnetic tape, a removable computer diskette, a random-access memory (RAM), a read-only memory (ROM), a rigid magnetic disk and an optical disk. Current examples of optical disks include compact disk-read only memory (CD-ROM), compact disk-read / write (CD-R / W) and DVD.

[0097] A data processing system suitable for storing and / or executing program code will include at least one processor (e.g., computer processor 32) coupled directly or indirectly to memory elements through a system bus. The memory elements can include local memory employed during actual execution of the program code, bulk storage, and cache memories which provide temporary storage of at least some program code in order to reduce the number of times code must be retrieved from bulk storage during execution. The system can read the inventive instructions on the program storage devices and follow these instructions to execute the methodology of the embodiments of the invention.

[0098] Network adapters may be coupled to the processor to enable the processor to become coupled to other processors or remote printers or storage devices through intervening private or public networks. Modems, cable modem and Ethernet cards are just a few of the currently available types of network adapters.

[0099] Computer program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object- oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the C programming language or similar programming languages.

[0100] It will be understood that blocks of the flowcharts shown in the figures and combinations of blocks in the flowcharts, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer (e.g., computer processor 32) or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowcharts and / or algorithms described in the present application. These computer program instructions may also be stored in a computer- readable medium (e.g., a non-transitory computer-readable medium) that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable medium produce an article of manufacture including instruction means which implement the function / act specified in the flowchart blocks and algorithms. The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowcharts and / or algorithms described in the present application.

[0101] Computer processor 32 is typically a hardware device programmed with computer program instructions to produce a special purpose computer. For example, when programmed to perform the algorithms described with reference to the figures, computer processor 32 typically acts as a special purpose diagnostics computer processor. Typically, the operations described herein that are performed by computer processor 32 transform the physical state of a memory, which is a real physical article, to have a different magnetic polarity, electrical charge, or the like depending on the technology of the memory that is used. For some embodiments, operations that are described as being performed by computer processor 32 are performed by a plurality of computer processors in combination with each other.

[0102] It will be appreciated by persons skilled in the art that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and subcombinations of the various features described hereinabove, as well as variations and modifications thereof that are not in the prior art, which would occur to persons skilled in the art upon reading the foregoing description.

Claims

CLAIMS1. A system, comprising: multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; and a processor, configured to: receive the signals, by processing the signals, identify a subset, of multiple subtypes of PD, to which the particular subtype belongs, and output an output indicating the subset of the subtypes.

2. The system according to claim 1, wherein the processor is configured to identify the subset by ascertaining whether the particular subtype is a rapid disease progression (RDP) subtype, and wherein the output indicates whether the particular subtype is the RDP subtype.

3. The system according to claim 1, wherein the processor is configured to identify the subset by ascertaining whether the particular subtype is a non-tremor dominant (NTD) subtype, and wherein the output indicates whether the particular subtype is the NTD subtype.

4. The system according to claim 1, wherein at least some of the signals were produced in response to magnetic stimulation of an occipital cortex of the brain.

5. The system according to claim 1, wherein the processor is further configured to compute, by processing the signals, an estimated PD duration for the patient, and wherein the output indicates the estimated duration.

6. The system according to claim 1, wherein the processor is further configured to compute, by processing the signals, an estimated PD stage of progression for the patient, and wherein the output indicates the estimated stage of progression.

7. The system according to any one of claims 1-6, wherein the processor is configured to process the signals by: computing one or more measures of neurophysiological activity exhibited in the signals, and inputting the measures of neurophysiological activity to a model.

8. The system according to claim 7, wherein the model includes a logistic regression model.

9. The system according to claim 7, wherein the measures of neurophysiological activity include a latency measure, and wherein the processor is configured to compute the latency measure by: computing at least one average signal, by averaging at least some of the signals, identifying a timing of a greatest local maximum of the average signal within a predefined time window, and computing the latency measure based on the timing.

10. The system according to claim 7, wherein the measures of neurophysiological activity include an interhemispheric connectivity measure, which quantifies a similarity between a right-side response of a right side of the brain to the magnetic stimulation and a left-side response of a left side of the brain to the magnetic stimulation.

11. A method, comprising: receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; by processing the signals, identifying, by the processor, a subset, of multiple subtypes of PD, to which the particular subtype belongs; and outputting an output indicating the subset of the subtypes.

12. The method according to claim 11, wherein identifying the subset comprises ascertaining whether the particular subtype is a rapid disease progression (RDP) subtype, and wherein the output indicates whether the particular subtype is the RDP subtype.

13. The method according to claim 11, wherein identifying the subset comprises ascertaining whether the particular subtype is a non-tremor dominant (NTD) subtype, and wherein the output indicates whether the particular subtype is the NTD subtype.

14. The method according to claim 11, wherein at least some of the signals were produced in response to magnetic stimulation of an occipital cortex of the brain.

15. The method according to claim 11, further comprising, by processing the signals, computing, by the processor, an estimated PD duration for the patient, wherein the output indicates the estimated duration.

16. The method according to claim 11, further comprising, by processing the signals, computing, by the processor, an estimated PD stage of progression for the patient, wherein the output indicates the estimated stage of progression.

17. The method according to any one of claims 11-16, wherein processing the signals comprises: computing one or more measures of neurophysiological activity exhibited in the signals; and inputting the measures of neurophysiological activity to a model.

18. The method according to claim 17, wherein the model includes a logistic regression model.

19. The method according to claim 17, wherein the measures of neurophysiological activity include a latency measure, and wherein computing the latency measure comprises: computing at least one average signal, by averaging at least some of the signals; identifying a timing of a greatest local maximum of the average signal within a predefined time window; and computing the latency measure based on the timing.

20. The method according to claim 17, wherein the measures of neurophysiological activity include an interhemispheric connectivity measure, which quantifies a similarity between a right-side response of a right side of the brain to the magnetic stimulation and a left-side response of a left side of the brain to the magnetic stimulation.

21. A computer software product comprising a tangible non-transitory computer-readable medium in which program instructions are stored, which instructions, when read by a processor, cause the processor to: receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from a particular subtype of Parkinson’s Disease (PD), in response to magnetic stimulation of the brain, by processing the signals, identify a subset, of multiple subtypes of PD, to which the particular subtype belongs, andoutput an output indicating the subset of the subtypes.

22. The computer software product according to claim 21, wherein the instructions cause the processor to identify the subset by ascertaining whether the particular subtype is a rapid disease progression (RDP) subtype, and wherein the output indicates whether the particular subtype is the RDP subtype.

23. The computer software product according to claim 21, wherein the instructions cause the processor to identify the subset by ascertaining whether the particular subtype is a non-tremor dominant (NTD) subtype, and wherein the output indicates whether the particular subtype is the NTD subtype.

24. A system, comprising: multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; and a processor, configured to: receive the signals, by processing the signals, compute an estimated PD duration for the patient, and output an output indicating the estimated PD duration.

25. A method, comprising: receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; by processing the signals, computing an estimated PD duration for the patient; and outputting an output indicating the estimated PD duration.

26. A computer software product comprising a tangible non-transitory computer-readable medium in which program instructions are stored, which instructions, when read by a processor, cause the processor to: receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain, by processing the signals, compute an estimated PD duration for the patient, and output an output indicating the estimated PD duration.

27. A system, comprising: multiple electrodes configured to record respective signals produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; and a processor, configured to: receive the signals, by processing the signals, compute an estimated PD stage of progression for the patient, and output an output indicating the estimated PD stage of progression.

28. A method, comprising: receiving, by a processor, multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering from Parkinson’s Disease (PD), in response to magnetic stimulation of the brain; by processing the signals, computing an estimated PD stage of progression for the patient; and outputting an output indicating the estimated PD stage of progression.

29. A computer software product comprising a tangible non-transitory computer-readable medium in which program instructions are stored, which instructions, when read by a processor, cause the processor to: receive multiple signals recorded by respective electrodes, the signals having been produced, by a brain of a patient suffering fromParkinson’s Disease (PD), in response to magnetic stimulation of the brain, by processing the signals, compute an estimated PD stage of progression for the patient, and output an output indicating the estimated PD stage of progression.

Images