WO2025101841A1 - Methods and apparatus for frontal temporal dementia diagnosis using a resting-state scalp eeg marker of regional interactions in the brain - Google Patents

Abstract

Examples may provide a method of determining a neurological disorder status of a subject. The method includes generating at least one dynamic brain network model (DNM) for the subject, and determining at least one sink index (SI) for the subject using the DNM. The method also includes determining whether the subject has frontotemporal dementia (FTD), Alzheimer's disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions. Additional methods as well as related systems and computer readable media are also provided.

Description

METHODS AND APPARATUS FOR FRONTAL TEMPORAL DEMENTIA DIAGNOSIS USING A RESTING-STATE SCALP EEG MARKER OF REGIONAL INTERACTIONS IN THE BRAIN

Cross-Reference to Related Applications

[0001] This application claims priority to U.S. Provisional Patent Application Ser. No. 63/597,001 , filed November 8, 2023, the disclosure of which is incorporated herein by reference.

Field

[0002] This disclosure relates generally to neurological disorder diagnostics.

Background

[0003] Frontotemporal dementia (FTD) is a rare neurological disorder that affects the prefrontal and anterior temporal cortex, leading to unusual behaviors, emotional problems, communication difficulties, and mobility issues. It represents 12- 20% of dementia cases and commonly manifests in individuals aged 45 to 64. For many years, research focused on identifying proteins or substances in the blood or cerebrospinal fluid, to monitor disease progression and assess treatment effectiveness. More recently, scientists are investigating ways to enhance brain imaging techniques and neuropsychological testing. Electroencephalogram (EEG) has emerged as a cost-effective and easily accessible tool for diagnosing and categorizing the seventy of dementia. Developing EEG biomarkers to complement neuropsychological tests for FTD detection is a desirable choice over other neuroimaging devices due to its low cost and accessibility.

[0004] Accordingly, there is a need for additional approaches to diagnose FTD in patients.

Summary

[0005] The present disclosure provides, in certain aspects, cost-effective and easily accessible tools for diagnosing and categorizing frontotemporal dementia, among other types of dementia. In some embodiments, for example, the present disclosure provides apparatus, systems, methods, and related standalone or bundled software for analyzing scalp EEG of dementia patients in hospitals, urgent care and other government-sponsored healthcare facilities. These and other aspects will be apparent upon a complete review of the present disclosure, including the accompanying figures.

[0006] According to various embodiments, a method of determining a neurological disorder status of a subject is presented. The method includes generating at least one dynamic brain network model (DNM) for the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions, thereby determining the neurological disorder status of the subject.

[0007] Various optional features of the above embodiments include the following. The method is at least partially computer-implemented. The method comprises generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject. The method comprises determining an SI heatmap for the subject. The SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject. The SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes. The method further comprises generating a report indicating the neurological disorder status of the subject. The method further comprises providing the neurological disorder status of the subject to the subject and/or to a healthcare provider. The method comprises administering at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease. The method comprises discontinuing administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

[0008] According to various embodiments, a system for determining a neurological disorder status of a subject is presented. The system includes an apparatus configured to obtain a scalp electroencephalogram (EEG) data set from the subject; and a controller operably connected to the apparatus, which controller comprises a processor, and a memory communicatively directly or remotely coupled to the processor, the memory storing non-transitory computer executable instructions which, when executed on the processor, perform operations comprising: generating at least one dynamic brain network model (DNM) for the subject using the scalp EEG data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

[0009] Various optional features of the above embodiments include the following. A wearable device comprises the apparatus. The non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject. The SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject. The SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes. The non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject. The report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease. The report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

[0010] According to various embodiments, a computer readable media is presented. The computer readable media comprises non-transitory computer executable instructions which, when executed by at least one electronic processor, perform at least: generating at least one dynamic brain network model (DNM) for a subject using a scalp electroencephalogram (EEG) data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

[0011] Various optional features of the above embodiments include the following. The non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject. The non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject. The SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject. The SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes. The non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject. The report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease. The report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

Drawings

[0012] The above and/or other aspects and advantages will become more apparent and more readily appreciated from the following detailed description of examples, taken in conjunction with the accompanying drawings, in which:

[0013] Fig. 1 is a flow chart that schematically shows exemplary method steps of determining a neurological disorder status of a subject according to some aspects disclosed herein; and

[0014] Fig. 2 is a schematic diagram of an exemplary system suitable for use with certain aspects disclosed herein.

[0015] Figs. 3A-3G. Sink index analysis pipeline. The Sink-Index analysis pipeline includes generating a patient-specific DNM from a few minutes of a patient's scalp EEG, constructing a coordinate system out of the row and column sums for each channel of each of the N matrices, using these coordinates to calculate the Sink-Index, and, examining the average Sink-Index per regions of interest in particular, the frontal- temporal vs the central-parietal-occipital regions, visualizing with box plots and brain topographic maps. [0016] Figs. 4A-4C. Three patients Sink-Index examples. Patient 1 (top) is an FTD patient. Patient 2 (middle) is an AD patient, and patient 3 (bottom) is a HC subject.

(A) After pre-processing, we displayed only selected channels and showed each patient ten to 30 minutes of scalp EEG data. (B) Brain Topographic Maps of each patient using Sink-Index vectors across all matrices for each patient’s recording. (C) Box plots of the average Sink-Index (across the entire recording) for the frontal- temporal and the central-parietal-occipital regions. Pathological brain regions for each condition present higher sink indices.

[0017] Fig. 5. Brain Topographic Maps per Cohort: The Sink-Index captures the brain network inter-dynamics at play, emphasizing pathological brain regions. FTD and AD patients have higher sink indices in their corresponding frontal-temporal and central-parietal-occipital pathological brain regions. FTD, HC, and AD patients from left to right display average sink indices per region of interest.

[0018] Figs. 6A-6D. Cohort-level Sink Index and MMSE behavior: The average sink indices for the FT and CPO regions for each cohort (A) capture the patient's pathological brain region dynamics. HCs display closely related sink indices between FT and CPO regions. Population characteristics such as Sex (B) and Age (C) do not seem to bias the average sink indices. In contrast, the MMSE score as a function of the Sink-Index ratio (D) shows similar scores corresponding to highly dispersed (high and low) sink indices across both the AD and FTD cohorts.

[0019] Figs. 7A-7C. Box Plot and ROC for Classifier Feature: We used box plots to visually indicate the distribution of each feature result across the 25th, 50th, and 75th percentiles and to locate outliers in the left panels, applying the ratio between frontal-temporal and central-parietal occipital regions for each feature. We plot the best classifier algorithm ROC corresponding to each feature in the right panels.

[0020] Figs. 8A and 8B. Average AUC and Precision Performance Benchmarks: The figure depicts the Sink-Index AUC and Precision in red. AUC (A) and Precision

(B) panels show the relative performance of each feature. The Sink-Index AUC presented values of 96.87±2.54, while Precision presented 84.39±16.54. The second- best technique, frequency power, presented values of 63.16±19.78 and 57.58±2.12 for AUC and Precision, respectively. Definitions

[0021] In order for the present disclosure to be more readily understood, certain terms are first defined below. Additional definitions for the following terms and other terms may be set forth throughout the specification. If a definition of a term set forth below is inconsistent with a definition in an application or patent that is incorporated by reference, the definition set forth in this application should be used to understand the meaning of the term.

[0022] As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise. Thus, for example, a reference to “a method” includes one or more methods, and/or steps of the type described herein and/or which will become apparent to those persons skilled in the art upon reading this disclosure and so forth.

[0023] It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. Further, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. In describing and claiming the methods, systems, and computer readable media, the following terminology, and grammatical variants thereof, will be used in accordance with the definitions set forth below.

[0024] Classifier. As used herein, “classifier” generally refers to algorithm computer code that receives, as input, test data and produces, as output, a classification of the input data as belonging to one or another class.

[0025] Data set: As used herein, “data set” refers to a group or collection of information, values, or data points related to or associated with one or more objects, records, and/or variables. In some embodiments, a given data set is organized as, or included as part of, a matrix or tabular data structure. In some embodiments, a data set is encoded as a feature vector corresponding to a given object, record, and/or variable, such as a given test or reference subject. For example, a medical data set for a given subject can include one or more observed values of one or more variables associated with that subject.

[0026] Subject: As used herein, “subject” or “test subject” refers to an animal, such as a mammalian species (e.g., human) or avian (e.g., bird) species. More specifically, a subject can be a vertebrate, e.g., a mammal such as a mouse, a primate, a simian or a human. Animals include farm animals (e.g., production cattle, dairy cattle, poultry, horses, pigs, and the like), sport animals, and companion animals (e.g., pets or support animals). A subject can be a healthy individual, an individual that has or is suspected of having a disease or pathology or a predisposition to the disease or pathology, or an individual that is in need of therapy or suspected of needing therapy. The terms “individual” or “patient” are intended to be interchangeable with “subject.” A “reference subject” refers to a subject known to have or lack specific properties (e.g., a known pathology, such as melanoma and/or the like).

[0027] Value: As used herein, “value” generally refers to an entry in a data set that can be anything that characterizes the feature to which the value refers. This includes, without limitation, numbers, words or phrases, symbols (e.g., + or -) or degrees.

Description of the Embodiments

[0028] Reference will now be made in detail to example implementations. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the invention. The following description is, therefore, merely exemplary.

[0029] I. Introduction

[0030] In some aspects, the present disclosure provides methods, systems, and computer readable media that utilize patient-specific dynamic “brain” network models (DNMs) from minutes of scalp EEG data. Using these DNMs, the methods, systems, and computer readable media calculate a new scalp EEG biomarker known as the "sink index," which characterizes how each brain region (node) is influenced by other regions in the brain network. Examining the sink index ratios of the brain regions associated with frontotemporal dementia to the remaining regions of the brain, the methods, systems, and computer readable media detect frontotemporal dementia while differentiating from healthy controls and other dementias like Alzheimer’s disease. These and other attributes of the present disclosure will be apparent upon a complete review of this specification, including the accompanying figures. [0031] To illustrate, Fig. 1 is a flow chart that schematically shows exemplary method steps of determining a neurological disorder status of a subject according to some aspects disclosed herein. As shown, method 100 includes generating at least one dynamic brain network model (DNM) for the subject (step 102). Method 100 also includes determining at least one sink index (SI) for the subject using the DNM (step 104). In addition, method 100 also includes determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions to thereby determining the neurological disorder status of the subject (step 106). In some embodiments, method 100 includes comprising generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject. In some embodiments, method 100 includes determining an SI heatmap for the subject. In some embodiments, the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject. In some embodiments, the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

[0032] The present disclosure also provides various systems and computer program products or machine readable media. In some aspects, for example, the methods described herein are optionally performed or facilitated at least in part using systems, distributed computing hardware and applications (e.g., cloud computing services), electronic communication networks, communication interfaces, computer program products, machine readable media, electronic storage media, software (e.g., machine-executable code or logic instructions) and/or the like. To illustrate, FIG. 2 provides a schematic diagram of an exemplary system suitable for use with implementing at least aspects of the methods disclosed in this application. As shown, system 200 includes at least one controller or computer, e.g., server 202 (e.g., a search engine server), which includes processor 204 and memory, storage device, or memory component 206, and one or more other communication devices 214, 216, (e.g., client-side computer terminals, telephones, tablets, laptops, other mobile devices, etc. (e.g., for receiving imaging data sets or results, etc.) in communication with the remote server 202, through electronic communication network 212, such as the Internet or other internetwork. Communication devices 214, 216 typically include an electronic display (e.g., an internet enabled computer or the like) in communication with, e.g., server 202 computer over network 212 in which the electronic display comprises a user interface (e.g., a graphical user interface (GUI), a web-based user interface, and/or the like) for displaying results upon implementing the methods described herein. In certain aspects, communication networks also encompass the physical transfer of data from one location to another, for example, using a hard drive, thumb drive, or other data storage mechanism. System 200 also includes program product 208 (e.g., for determining a neurological disorder status of a subject as described herein) stored on a computer or machine readable medium, such as, for example, one or more of various types of memory, such as memory 206 of server 202, that is readable by the server 202, to facilitate, for example, a guided search application or other executable by one or more other communication devices, such as 214 (schematically shown as a desktop or personal computer). In some aspects, system 200 optionally also includes at least one database server, such as, for example, server 210 associated with an online website having data stored thereon (e.g., entries patient data sets, etc.) searchable either directly or through search engine server 202. System 200 optionally also includes one or more other servers positioned remotely from server 202, each of which are optionally associated with one or more database servers 210 located remotely or located local to each of the other servers. The other servers can beneficially provide service to geographically remote users and enhance geographically distributed operations.

[0033] As understood by those of ordinary skill in the art, memory 206 of the server 202 optionally includes volatile and/or nonvolatile memory including, for example, RAM, ROM, and magnetic or optical disks, among others. It is also understood by those of ordinary skill in the art that although illustrated as a single server, the illustrated configuration of server 202 is given only by way of example and that other types of servers or computers configured according to various other methodologies or architectures can also be used. Server 202 shown schematically in FIG. 2, represents a server or server cluster or server farm and is not limited to any individual physical server. The server site may be deployed as a server farm or server cluster managed by a server hosting provider. The number of servers and their architecture and configuration may be increased based on usage, demand and capacity requirements for the system 200. As also understood by those of ordinary skill in the art, other user communication devices 214, 216 in these aspects, for example, can be a laptop, desktop, tablet, personal digital assistant (PDA), cell phone, server, or other types of computers. As known and understood by those of ordinary skill in the art, network 212 can include an internet, intranet, a telecommunication network, an extranet, or world wide web of a plurality of computers/servers in communication with one or more other computers through a communication network, and/or portions of a local or other area network.

[0034] As further understood by those of ordinary skill in the art, exemplary program product or machine readable medium 208 is optionally in the form of microcode, programs, cloud computing format, routines, and/or symbolic languages that provide one or more sets of ordered operations that control the functioning of the hardware and direct its operation. Program product 208, according to an exemplary aspect, also need not reside in its entirety in volatile memory, but can be selectively loaded, as necessary, according to various methodologies as known and understood by those of ordinary skill in the art.

[0035] As further understood by those of ordinary skill in the art, the term "computer-readable medium" or “machine-readable medium” refers to any medium that participates in providing instructions to a processor for execution. To illustrate, the term "computer-readable medium" or “machine-readable medium” encompasses distribution media, cloud computing formats, intermediate storage media, execution memory of a computer, and any other medium or device capable of storing program product 208 implementing the functionality or processes of various aspects of the present disclosure, for example, for reading by a computer. A "computer-readable medium" or “machine-readable medium” may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media includes, for example, optical or magnetic disks. Volatile media includes dynamic memory, such as the main memory of a given system. Transmission media includes coaxial cables, copper wire and fiber optics, including the wires that comprise a bus. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications, among others. Exemplary forms of computer-readable media include a floppy disk, a flexible disk, hard disk, magnetic tape, a flash drive, or any other magnetic medium, a CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a RAM, a PROM, and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave, or any other medium from which a computer can read.

[0036] Program product 208 is optionally copied from the computer-readable medium to a hard disk or a similar intermediate storage medium. When program product 208, or portions thereof, are to be run, it is optionally loaded from their distribution medium, their intermediate storage medium, or the like into the execution memory of one or more computers, configuring the computer(s) to act in accordance with the functionality or method of various aspects disclosed herein. All such operations are well known to those of ordinary skill in the art of, for example, computer systems.

[0037] In some aspects, program product 208 includes non-transitory computer-executable instructions which, when executed by electronic processor 604, perform at least: generating at least one dynamic brain network model (DNM) for a subject using a scalp electroencephalogram (EEG) data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

[0038] Typically, a scalp EEG data set is obtained from subject 220 using apparatus (e.g., an EEG sensor headset) 218.

[0039] I. Description of Example Embodiments

[0040] Example: Sink Index: A Network Based EEG Biomarker for

Frontotemporal Dementia and Alzheimer’s Disease

[0041] Frontotemporal dementia (FTD) is a clinically complex neurodegenerative illness characterized by progressive deteriorations in temperament, judgment, conduct, and communication. FTD is a prevalent condition in adults below 60 years of age diagnosed with dementia, and it is the primary cause of cognitive decline, alongside Alzheimer's disease (AD). Diagnosis requires careful clinical interview and examination to identify the syndrome by describing the principal characteristics, chronology, and tempo that define the syndrome, and radiological profiles provide crucial support. Misdiagnosis and late diagnosis are not uncommon, owing to the complexity of the phenotypes overlaps of features with those of other neurodegenerative syndromes and psychiatric disorders. Furthermore, the preclinical features have not yet been established — confounding diagnosis in those cases that have not yet developed a typical syndrome.

[0042] Medical professionals diagnose dementia-related disorders using clinical examinations and advanced neurological imaging tools such as FDG-PET, SPECT, and MRI. These methods help identify abnormalities in the brain structure and distinct patterns of neurodegeneration characteristic of each condition. AD predominantly affects the hippocampus and certain temporal and parietal neocortex areas, while FTD affects the frontal and anterior temporal regions. There needs to be more other biomarkers for early diagnosis. Serum neurofilament light chain levels can predict which carriers of causal genetic mutations will develop an FTD prodrome. However, NfL levels vary depending on genetic and clinical factors, and NfL elevation occurs in other neurodegenerative diseases — it is not specific to FTD. Generally, these techniques come into play only after significant neurodegeneration has occurred and dementia has become established.

[0043] While MRI, SPECT, and PET are the tools of the trade for diagnosing these disorders by assessing brain activity and connectivity patterns, their worldwide adoption remains debatable as scans are expensive and often inaccessible, especially to patients in underdeveloped countries. There is a need for a diagnostic test that detects early physiological changes in dementia, distinguishes FTD from other neurodegenerative conditions such as AD and is widely accessible. A viable alternative could be scalp electroencephalography (EEG), an established, widely available, low-cost technology that provides an intriguing alternative; however, a utility for early diagnosis requires EEG markers that discriminate between dementia syndromes. Researchers have yet to identify reliable EEG markers for AD and FTD. Both disorders underscore the critical requirement for a non-invasive, cost-effective, universally accessible early diagnostic method.

[0044] Over recent decades, researchers have made significant strides in utilizing EEG data to study AD and FTD through various signal analysis techniques. In the time domain, researchers have documented a decrease in signal complexity in EEG readings of AD patients by using Approximate Entropy (ApEn) and Sample Entropy (SampEn). In the time domain, studies employing connectivity features and SVM to classify AD+FTD/HC or AD/FTD achieved a modest 73% accuracy. In the frequency domain, studies employing Power Spectrum analysis in EEG rhythms and AD/FTD/HC classification algorithms have achieved 85-93% accuracy rates. However, these studies use small datasets (less than 40 subjects) and only tackle common classification problems of FTD/HC or AD/HC but not necessarily all three. Similarly, in the time-frequency domain, several studies using Discrete Fourier Transform and Wavelets, combined with classification algorithms for AD/FTD/HC, reported accuracy rates ranging from 83-86%. As in the frequency domain studies, these time frequency domain studies lack large datasets or miss FTD or AD subjects as part of their population. Furthermore, these techniques use individual channel features or pairwise and static connectivity features, which cannot capture the n-to-n regional dynamics across the brain network. Despite these advancements, the clinical validation of an EEG biomarker capable of precisely detecting AD and FTD remains elusive, which is vital for ensuring early intervention and improved care for affected individuals.

[0045] In this study, we leverage EEG recordings to construct patient-specific Dynamic Network Models (DNM) to derive a computational biomarker for dementia detection. These DNMs capture brain regions' spatial interactions and temporal influences on the brain network, thereby identifying pathological network nodes (EEG channels). We hypothesize that these pathological network nodes do not influence other brain areas while other brain regions attempt to compensate for their reduced activity. To test our hypothesis, we employed a novel scalp EEG biomarker called the Sink-Index that captures the dynamics of nodes in a patient's brain network, connectivity properties, and the interplay among various brain regions. Utilizing a publicly available scalp EEG dataset, which includes recordings from AD, FTD, and HC subjects, we applied our analysis pipeline, which involved creating a patientspecific DNM to calculate our metric across all brain regions. We compared the biomarker's output for the combined Frontal-Temporal versus Central-Parietal- Occipital regions and incorporated the ratio of these as a feature into several supervised learning algorithms. Accuracy results, measured in terms of the Area Under the Curve (AUC) and Precision, which are standard measures of classification performance, obtained from a leave-one-out cross-validation method suggest that patient-specific DNMs combined with the Sink-Index EEG biomarker outperform traditional spectral and time-frequency analysis. These findings suggest a promising, accurate, and cost-effective alternative for dementia detection.

[0046] Materials and methods

[0047] Dataset Description

[0048] We downloaded and analyzed a dataset from the OpenNeuro website containing EEG recordings from AD, FTD, and HC subjects. The 2nd Department of Neurology at AHEPA General Hospital in Thessaloniki initially recorded this dataset, which included eighty-eight (88) subjects. Among these, clinicians diagnosed thirty-six (36) with AD, twenty-three (23) with FTD, and identified twenty-nine (29) as healthy controls (HC). Notably, none of the subjects reported any comorbidities. Table 1 summarizes the patient population statistics.

Table 1

Dataset demographic: The statistical values in parenthesis represents standard deviation for each quantity of each cohort

[0049] Regarding the demographics of the study groups, the average ages were 66.4 (±7.9) years for the AD group, 67.8 (±5.4) years for the FTD group, and 63.7 (±8.2) years for the HC group. The Mini-Mental State Examination (MMSE) scores varied significantly among the groups. The AD group had an average score of 17.75 (±4.5), indicating a moderate level of cognitive impairment. The HC group presented with an average score of 30 (±0), reflecting normal cognitive function. In contrast, the FTD group showed an average MMSE score of 22.1 (±2.6), suggesting mild cognitive impairment.

[0050] Data Acquisition and Preprocessing [0051] For creating the recordings, the clinical team at the AHEPA General Hospital used the Nihon Kohden EEG 2100 device configured in a standard 10-20 montage and operated at a sampling rate of 500 Hz, with a resolution of 10pV/mm. During the recording sessions, subjects were seated in a relaxed posture with closed eyes. We down-sampled all recordings to 250 Hz, filtered them using Butterworth band-pass filter (BPF) 0.5 - 48 Hz, and notch filtered them at 50 Hz and their harmonics with a stopband of 2 Hz to remove power line noise interference. We used MATLAB R2023b Update 7 (MathWorks, 2024) to process and analyze the data. We built models, Receiver Operating Curves (ROCs), AUCs, and Confusion Matrices using Python 3.12.0 (Python Software Foundation, Wilmington, DE).

[0052] Dynamic network models

[0053] DNMs are a class of mathematical generative models that describe the evolution of complex systems composed of interconnected components. These models capture how the network structure of the system changes over time in response to various internal and external factors. Specifically, in the case of EEG signals, DNMs characterize how each channel dynamically influences the rest of the brain network. Our study constructed Linear Time-Varying (LTV) DNMs for each subject based on their EEG recordings by concatenating a sequence of discrete time Linear Time-Invariant (LTI) models. Each model was calculated over a specified time window, T, using the state evolution equation: x(t +1 ) = Ax(t) (1 )

In this equation, x(t) e R^N, symbolizes the EEG channels, and A e R^NxN is the state evolution matrix. The state transition matrix A describes the channel interactions and influences of channels over time. The sum of all T windows amounts to the size of the recording. To obtain each patient-specific DNM, we applied least square estimation techniques using the first-order multivariate autoregressive model expressed in (1 ). By employing this method, we were able to accurately capture the dynamics of the EEG channels' interactions, providing deeper insights into the brain's network behavior.

[0054] Sink Index Metric

[0055] In 2022, others used intracranial EEG (iEEG) data to create patientspecific LTV DNMs using equation (1 ) from iEEG electrode signals. 26 The authors conceptualized iEEG channels or iEEG nodes as "sources" and "sinks" in the brain network. In their approach, "sources" denote brain regions that exert significant influence over other nodes without being influenced by the network. Conversely, "sinks" are regions predominantly influenced by other nodes without exerting much influence themselves.

[0056] Furthermore, they constructed a normalized Cartesian coordinate system using each channel's row rank and column rank in the state evolution matrix as abscissa and ordinate coordinates, respectively. Specifically, the following two norms were computed and then ranked and normalized such that the largest rr and cr value was 1 , and the smallest value was 1/N.

The above values are the norms prior to ranking and normalization. Leveraging this coordinate system, the authors introduced a metric to determine the proximity of a node to the ideal sink (pathological brain circuit) located at coordinates (1 ,1/N), where N denotes the number of channels. They labeled this metric Sink-Index as depicted in equation (3).

With this understanding of the source-sink theory, generating patient-specific DNM and calculating the Sink-Index for all EEG channels might indicate that frontal and temporal brain regions in FTD act as strong sinks. Consequently, the Sink-Index distribution would be more heterogeneous for AD, and these diseases would differentiate from HCs with no solid sources or sinks.

[0057] Analysis Pipeline

[0058] Figure 3 depicts the analysis pipeline used for the Sink-Index portion of this study. For each subject scalp EEG recording (Figure 3A), we separated the recording into smaller intervals of time duration T (Figure 3B). We calculated the state evolution matrix for each interval T using least square estimation (Figure 3C). For each matrix k, we compute all row and column rank coordinate pairs (rr, cr) for each channel as described above and shown in Figure 3D. We then substituted the coordinates for each channel into equation (3) and found the Sink-Index for channel i of matrix k. We repeated the process for N channels and M matrices, resulting in a Sink-Index vector of size N per time, which we plotted into a heatmap where the heat is the Sink-Index as calculated in Equation 2 (Figure 3E). We grouped the channels into two groups, namely, the frontal-temporal group (Fp1 , Fp2, F3, F4, F7, F8, T3, and T4) and the central parietal-occipital group (01 , T5, 02, T6, Cz, Pz, P3, Fz, P4, C3, and C4). For each group of channels, we calculated the average Sink-Index across the entire group and box plotted both FT and CPO groups per subject, per cohort, across cohorts, and as a Sink-Index ratio FT/CPO across all subject cohorts (AD, FTD, and HC) (Figure 3F). Finally, we displayed the Sink Index as a brain topographic map, as shown in Figure 3G.

[0059] Spectral and Time-Frequency Analysis

[0060] We perform the Sink index analysis in the time domain from patientspecific DNMs. This study also analyzed the data using frequency and time-frequency techniques for comparison and validation. In the frequency domain, we performed an average frequency power analysis. We decomposed the signal into functionally distinct frequencies at (1-4 Hz), (4-8 Hz), (8-12 Hz), and (12-30 Hz) corresponding to the delta, theta, alpha, and beta bands. We calculated the Fast-Fourier transform for each band and squared the magnitude to obtain the power spectrum. We averaged the power spectrum across the entire recording for frontal-temporal and central parietal-groups and box plotted both FT and CPO groups per subject and cohort across cohorts.

[0061] The time-frequency analysis used wavelet decomposition and Hjorth parameters to extract samples by taking eight-second windows with fifty (50) percent successive overlap. Five-level dyadic wavelet decomposition (using 'sym4', 'coif2', 'haar,' and 'db4' templates) helped capture the data's temporal and spectral features across all channels. At the same time, Hjorth parameters (activity, mobility, and complexity) provided complementary measures of the neural activity. Principal Components Analysis (PCA) provided feature reduction from an initial 285 features averaged across all 8-second windows to 5 features per subject.

[0062] Feature Extraction and Classification

[0063] Supervised learning algorithms provide a mechanism to predict and infer relationships between data features and targets. We contrasted the accuracy, specificity, and sensitivity of seven (7) different algorithms, namely (linear discriminant analysis, quadratic discriminant, k-nearest neighbor, naive Bayes, decision trees, random forest, logistic regression, and SVM) to select the best classifier for AD, FTD, and HC. The Sink-Index ratio, frequency band energy, and Hjorth parameters provided the features of the classification algorithms for each of the signal analysis techniques used in the time, frequency, and time-frequency domain. The classification problem addressed was multiclass using the One-vs-the-Rest approach, which compares the positive class against the combination of the remaining classes, considered as aggregate, the negative class and assumed to be one. This approach tested the ability of the features to classify in the presence of all conditions correctly. We used three distinct features for classification based on the signal analysis technique. We used the ratio derived from the averaged Frontal Temporal Nodes Sink Index divided by the averaged combined Central-Parietal-Occipital Nodes Sink Index for the Sink-Index analysis. For the frequency analysis, the power spectrum across the entire recording for frontal-temporal and central-parietal-occipital groups served as features, while in time-frequency analysis, the PCA-reduced Hjorth parameters and wavelet coefficients provided the corresponding classifying feature needed by the algorithms.

[0064] Statistical analysis

[0065] Since our dataset was relatively small, less than 100 patients, we used a leave-one-patient-out cross-validation procedure to compute a reliable, unbiased, and accurate estimate of the performance of our models. In each split, the hyperparameter k, which controls the number of subsets of the data, was set to two, namely, one training data set containing all but one patient (87 patients) and one test dataset containing the testing patient. We performed this task 88 times. We used a Receiver Operating Characteristic (ROC) metric to evaluate the quality of multiclass classifiers, plotting binarized true positive rates (TPR) against false positive rates (FPR). To determine the contributions from all the classes, we micro-averaged both rates using equations (3) and (4), respectively. For TPR, we divided the sum of all true positive (TP) cases overall classes C by the sum of TPs and false negative (F/V) cases overall classes C. Similarly, for FPR, we divide the sum of all false positive (FP) cases overall classes C by the sum of FPs and true negative (T/V) cases overall classes C. We used the area under the curve to measure model performance: accuracy, precision, sensitivity, and specificity. We compared the performance metrics of each analysis technique using the AUC and Precision.

[0066] Results

[0067] The Sink Index highlights pathological regions for FTD and AD patients

[0068] Averaging the Sink-Index over the entire recording gave us one data point per channel per subject. Figure 4 illustrates the scalp EEG, brain topographic map, and average Sink-Index across all channels grouped by frontal-temporal and central-parietal-occipital regions for three subjects with FTD (Patient 1 ), AD (Patient 2), and HC status (Patient 3). In the FTD case, Patient 1 is a 57-year-old female with an MMSE score of 22. In this patient, the frontal temporal nodes display higher sink indices than the central-parietal-occipital nodes corresponding to the pathophysiology of the FTD disorder. The brain topographic map shows higher sink indices in the Fp1 and Fp2 nodes corresponding to the prefrontal regions. Likewise, T3 and T4 nodes in the frontal-temporal region present higher sink indices than those in the central- parietal-occipital region. The box plot confirms the brain topographic map results.

[0069] AD Patient 2 is a 70-year-old male with an MMSE score of 14. In this patient, the frontal temporal nodes display lower sink indices than the central-parietal- occipital nodes corresponding to the pathophysiology of the AD disorder. The topographic map of patient sub-03’s brain shows higher sink indices in the F7, C3, Cz, C4, and P8 nodes corresponding to the frontal, central, and parietal regions. The box plot confirms higher sink indices in the central parietal-occipital regions.

[0070] HC Patient 3 is a 67-year-old female with an MMSE score of 30. This patient presents balanced average sink indices for frontal-temporal and central- parietal-occipital regions. This patient displays relatively low sink indices in topographic brain maps and similar numbers in the box plot of the subject’s brain.

[0071] Common cohort-level characteristics emerge within FTD, AD, and HCs

[0072] Figure 5 depicts the topographic maps of the brains for each cohort. The topographic maps and sink indices reveal that FTD and HC cohorts exhibit more homogeneous patterns, with FTD patients showing concentrated activity in the frontal- temporal regions and HCs displaying a balanced distribution of Sink-Index magnitudes across the brain. In contrast, the AD cohort demonstrates a more heterogeneous pattern, with higher sink indices in the central-parietal occipital regions. Some AD patients also exhibit high indices in the frontal-temporal regions, corresponding to the disease's pathophysiology. Advanced AD cases with high Tau concentrations may show frontal-temporal dysfunction and atrophy, corresponding to symptoms of altered behavior and personality changes. Nevertheless, the Sink-Index characterizes these brain regions within each cohort.

[0073] Figure 6A summarizes our findings following the approach used to analyze patients 1 -3 from Figure 4. For each cohort patient, we captured the average Sink-Index for the entire recording for all channels in the frontal-temporal vs the central-occipital-parietal regions. We observed that the Sink-Index of the EEG electrodes associated to the frontal-temporal brain regions (Fp1 , Fp2, F3, F4, F7, F8, T3, T4) and central-parietal-occipital regions (01 , T5, 02, T6, Cz, Pz, P3, Fz, P4, C3, C4) display noticeable differences between FTD (1.3389 ± 0.0895 vs. 0.8444 ± 0.0651 ), ALZ (0.6015 ± 0.0188 vs. 0.7766 ± 0.0158) and HC (0.8978 ± 0.0453 vs. 0.9116 ± 0.0457). Several studies have reported significant differences in FTD and AD patients depending on sex and age, suggesting implications in pathogenesis and clinical features in the pathophysiology.

[0074] Panel B depicts the Sink-Index per cohort and Sex (Female or Male), while Panel C captures the Sink-Index behavior for all three cohorts per age group. We used five age groups, namely, 44-51 (group 1 ), 52-58 (group 2), 59-65 (group 3), 66-72 (group 4), and 73-79 (group 5) years old. Early onset AD (EOAD) patients are typically younger than 65 years old. The data set contains 16 patients of age less than 65. In the case of FTD, the data set contains one patient of age 44, considered early onset, since, in most cases, FTD occurs between the ages of 45 and 65. The Sink- Index identified those patients in groups 1 and 1 -3 for FTD and AD, respectively. Table 2 lists the early-onset participants and their demographics. Finally, on panel D of Figure 6, we plot the average Sink-Index of each patient against the patient's MMSE. Several patients from the FTD and AD cohorts had equal MMSE but different Sink- Index ratios (Sink-Index of frontal-temporal regions divided by Sink-Index of central- parietal-occipital regions). Similarly, the HC group all had MMSE scores 30 with a Sink- Index ratio spreading between .85 and 1.1.

Table 2

Early Onset Demographics: EO for FTD < 45 and EO for Alz < 65. Dataset contains

1 and 16 participants of each cohort, respectively.

Participant Gender Age Condition MMSE Age_Group sub-076 M 44 F 24 1 sub-028 M 49 A 20 1 sub-029 F 53 A 16 2 sub-030 F 56 A 20 2 sub-001 F 57 A 16 2 sub-035 F 57 A 22 2 sub-036 F 58 A 9 2 sub-032 F 59 A 20 3 sub-023 M 60 A 16 3 sub-006 F 61 A 14 3 sub-015 M 61 A 18 3 sub-017 F 61 A 6 3 sub-026 F 61 A 18 3 sub-008 M 62 A 16 3 sub-019 F 62 A 14 3 sub-012 M 63 A 18 3 sub-013 F 64 A 20 3

[0075] The Sink-Index Ratio is a robust classifier

[0076] The Sink-Index captures the dynamics occurring among the channels in the Fronto-Temporal Regions and among the channels in the Central, Parietal, and Occipital regions. To create a single classification feature, we divided the FT regions' Sink-Index by the CPO regions' Sink Index. We applied the same grouping and ratio concepts to the features provided by each analysis technique to obtain a single classifier. Namely, for the frequency analysis, we used the power ratio for frontal- temporal and central-parietal-occipital groups, and for the Hjorth Parameters, we used PCA-reduced Hjorth parameters, five parameters total.

[0077] Figure 7 depicts the feature distributions using box plots (left panels) and the best-performing multiclass ROCs (right panels) using the One vs the Rest multiclass approach. In panel 5A, the Sink-Index ratio box plot displays interquartile distributions for the FTD, HC, and ALZ cohorts with median Sink-Index ratio values of 1.24, .97, and .76, respectively. The box-plot summaries for each cohort show minimum to zero IQR overlap. The corresponding Sink index ratio ROCs display AUCs of .96, .99, and .95 for ALZ, FTD, and HC. In Panel 5B, the IQR for the HC cohort shows some overlap with FTD and ALZ for the Alpha band frequency power feature. However, between pathologies, IQR overlap was evident. The corresponding frequency power ROCs display AUCs of .69, .38, and .79 for ALZ, FTD, and HC. Conversely, in Panel 5C, the Hjorth parameter feature displays total IQR overlap across all cohorts. The corresponding Hjorth parameter ROCs display AUCs barely above or at chance levels of .50, .53, and .62 for ALZ, FTD, and HC.

[0078] Sink Index Ratio outperforms features from other analysis techniques

[0079] Random Forest performs best for the Sink-Index ratio compared to all other classification techniques. As depicted in Figure 7, on the lower panel left, the Sink-Index ratio ROC shows an AUC of 100% for FTD, 96% for AD, and 95% for HCs. For the power spectrum ratio, in Figure 7, on the lower panel right, SVM performs the best compared to all other classification techniques. However, the power spectrum ratio ROC for the Alpha frequency shows diminished performance with AUCs of 38% for FTD, 69% for AD, and 79% for HCs. Figure 8 depicts the average AUC and precision performance benchmarks per cohort, sorted by mean for each analysis technique. The AUC for Sink-Index across all cohorts presented values of 96.87±2.54, while the power spectrum ratio and Hjorth parameters resulted in 63.16±19.78 and 51 ,90±10.22, respectively. Panel A shows how the AUC of the Sink-Index is more than 12% higher for the Sink-Index ratio feature in HCs 95.25±1.28 versus the subsequent best representation, the power spectrum ratio 82.62±1.99. Precision measures how well the model is capable of identifying the positive class. We determine it by dividing the total correct positive predictions (true positives) by the sum of all predictions classified as positive, including both correct (true positives) and incorrect (false positives) predictions. In our multiclass classification problem, the positive class is one, and the negative class is the composition of the remaining two. Panel B of Figure 8 shows Precision per feature sorted by mean. The average Sink-Index Precision of 84.39±16.54 outperforms the next best feature, frequency power 57.58±2.12, by at least 27%. Once again, the Sink-Index ratio outperforms other features from other analysis techniques.

[0080] Discussion

[0081] This study represents a significant advancement in the diagnosis of Frontotemporal Dementia and Alzheimer's disease, two challenging neurodegenerative disorders with complex and overlapping pathologies. Amyloid-beta plaques and Tau protein tangles characterize AD, leading to neuronal death in areas like the hippocampus, which results in memory loss and cognitive decline. FTD involves degeneration in the frontal and temporal lobes, often linked to proteinopathies like Tau, TDP-43, and FUS proteins, leading to behavioral and language deficits. Traditional diagnostic techniques, such as MRI, PET, or SPECT, have limitations since they detect neurodegeneration only after it has become significant. This study introduces a novel EEG-based biomarker - the Sink Index, derived from patientspecific Dynamic Network Models (DNM). This approach provides a refined perspective to observe the intricate dynamics within the brain's neural network, offering a unique insight into the underlying brain activity associated with AD and FTD. [0082] The development of the Sink Index opens up new possibilities for the utilization of EEG data in the differential diagnosis of dementia. It shows promise in distinguishing between FTD and AD cases, as well as between neurodegenerative disease cases and controls, with strong classification properties relative to other EEG analysis approaches. Importantly, it provides information about cortical function comparable to that of FDG-PET in terms of dementia detection and diagnostically relevant profiles. While further validation work is necessary, the potential of the Sink Index for diagnostic-quality brain function mapping using inexpensive, widely available, and portable technology is substantial. Its capability for remote measurement and monitoring in both clinical practice and research, as well as its potential for use in resource limited environments, holds promise for enhancing dementia diagnosis and management.

[0083] Sink Index as a biomarker for FTD diagnosis

[0084] The Sink Index successfully highlighted pathological regions in AD and FTD patients, aligning with the known neurodegenerative patterns of these diseases. For AD, a lower Sink Index in frontal-temporal nodes compared to central-parietal- occipital nodes was observed, reflecting the synaptic dysfunction and neuronal loss in regions critical for memory and cognition. Conversely, a higher Sink Index in frontal- temporal nodes in FTD indicated the disorder's pathology, highlighting the atrophy and neuronal loss in areas responsible for behavior and language.

[0085] The Sink Index Ratio emerged as a robust classifier, distinguishing between AD, FTD, and HCs with high accuracy, as evidenced by the high Area Under the Curve (AUC) values. This novel metric offers diagnostic potential and sheds light on the compensatory mechanisms in the brain. In AD, for instance, the altered network dynamics might reflect the brain's attempt to maintain functional connectivity despite neuronal loss, while in FTD, the increased Sink Index in frontal-temporal nodes could indicate a heightened reliance on these regions due to loss of function in others. These insights are crucial as they mirror the complex neurobiological processes in neurodegenerative diseases.

[0086] Sink-Index Biology for FTD Detection

[0087] The biological implications of our findings extend beyond diagnostics. Tau-PET imaging highlights the clinically relevant biochemistry of Tau, displaying functional changes in brain glucose metabolism; MRI assesses vascular changes and subtle patterns of atrophy. At the same time, Sink-Index tracks the influence (or lack thereof) of brain regions on other regions of the brain network through time by calculating a patient-specific DNM from a Scalp EEG, providing a cost-effective and reliable alternative for FTD diagnosing. The Sink Index provides a window into the intricate neural interactions and compensatory mechanisms vital to understanding neurodegenerative diseases by capturing dynamic interactions within the brain's neural network. Early-onset FTD and AD detection remains a challenge in today's clinical arena. The early appearance of pathological conditions and the progressive nature of these dementias emphasized the need for biomarkers sensitive to the brain changes expected prior to the onset of clinical symptoms. The Sink-Index may provide an alternative for early-onset diagnosis of FTD and AD. Grasping this concept is crucial for creating specific interventions and management plans that could slow the progression of the disease.

[0088] In conclusion, our research introduces a novel diagnostic tool for AD and FTD and contributes significantly to the broader understanding of these conditions. Through its portrayal of neural network dynamics, the Sink-Index holds promise as a non-invasive, cost-effective diagnostic tool. Its effectiveness in identifying the unique neurodegenerative patterns of AD and FTD highlights its potential to transform the diagnosis and management of these complex conditions.

[0089] Some further aspects are defined in the following clauses:

[0090] Clause 1 : A method of determining a neurological disorder status of a subject, the method comprising: generating at least one dynamic brain network model (DNM) for the subject; determining at least one sink index (SI) for the subject using the DNM; and, determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions, thereby determining the neurological disorder status of the subject.

[0091] Clause 2: The method of Clause 1 , comprising generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject.

[0092] Clause 3: The method of Clause 1 or Clause 2, comprising determining an SI heatmap for the subject.

[0093] Clause 4: The method of any one of the preceding Clauses 1 -3, wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

[0094] Clause 5: The method of any one of the preceding Clauses 1 -4, wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

[0095] Clause 6: The method of any one of the preceding Clauses 1-5, further comprising generating a report indicating the neurological disorder status of the subject.

[0096] Clause 7: The method of any one of the preceding Clauses 1-6, further comprising providing the neurological disorder status of the subject to the subject and/or to a healthcare provider.

[0097] Clause 8: The method of any one of the preceding Clauses 1-7, comprising administering at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease. [0098] Clause 9: The method of any one of the preceding Clauses 1 -8, comprising discontinuing administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

[0099] Clause 10: A system for determining a neurological disorder status of a subject, the system comprising: an apparatus configured to obtain a scalp electroencephalogram (EEG) data set from the subject; and, a controller operably connected to the apparatus, which controller comprises a processor, and a memory communicatively directly or remotely coupled to the processor, the memory storing non-transitory computer executable instructions which, when executed on the processor, perform operations comprising: generating at least one dynamic brain network model (DNM) for the subject using the scalp EEG data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

[00100] Clause 11 : The system of Clause 10, wherein a wearable device comprises the apparatus.

[00101] Clause 12: The system of Clause 10 or Clause 11 , wherein the non- transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject.

[00102] Clause 13: The system of any one of the preceding Clauses 10-12, wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

[00103] Clause 14: The system of any one of the preceding Clauses 10-13, wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

[00104] Clause 15: The system of any one of the preceding Clauses 10-14, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject.

[00105] Clause 16: The system of any one of the preceding Clauses 10-15, wherein the report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease.

[00106] Clause 17: The system of any one of the preceding Clauses 10-16, wherein the report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

[00107] Clause 18: A computer readable media, comprising non-transitory computer executable instructions which, when executed by a processor, perform operations comprising: generating at least one dynamic brain network model (DNM) for a subject using a scalp electroencephalogram (EEG) data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

[00108] Clause 19: The computer readable media of Clause 18, wherein the non- transitory computer executable instructions which, when executed by the processor, further perform operations comprising: generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject.

[00109] Clause 20: The computer readable media of Clause 18 or Clause 19, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject.

[00110] Clause 21 : The computer readable media of any one of the preceding Clauses 18-20, wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

[00111] Clause 22: The computer readable media of any one of the preceding Clauses 18-21 , wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

[00112] Clause 23: The computer readable media of any one of the preceding Clauses 18-22, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject.

[00113] Clause 24: The computer readable media of any one of the preceding Clauses 18-23, wherein the report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease.

[00114] Clause 25: The computer readable media of any one of the preceding Clauses 18-24, wherein the report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

[00115] While the invention has been described with reference to the exemplary embodiments thereof, those skilled in the art will be able to make various modifications to the described embodiments without departing from the true spirit and scope. The terms and descriptions used herein are set forth by way of illustration only and are not meant as limitations. In particular, although the method has been described by examples, the steps of the method can be performed in a different order than illustrated or simultaneously. Those skilled in the art will recognize that these and other variations are possible within the spirit and scope as defined in the following claims and their equivalents. All patents, patent applications, other publications or documents, and the like cited herein are incorporated by reference in their entirety for all purposes to the same extent as if each individual item were specifically and individually indicated to be so incorporated by reference.

Claims

What is claimed is:

1 . A method of determining a neurological disorder status of a subject, the method comprising: generating at least one dynamic brain network model (DNM) for the subject; determining at least one sink index (SI) for the subject using the DNM; and, determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions, thereby determining the neurological disorder status of the subject.

2. The method of claim 1 , comprising generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject.

3. The method of claim 1 , comprising determining an SI heatmap for the subject.

4. The method of claim 1 , wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

5. The method of claim 1 , wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

6. The method of claim 1 , further comprising generating a report indicating the neurological disorder status of the subject.

7. The method of claim 1 , further comprising providing the neurological disorder status of the subject to the subject and/or to a healthcare provider.

8. The method of claim 1 , comprising administering at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease.

9. The method of claim 1 , comprising discontinuing administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

10. A system for determining a neurological disorder status of a subject, the system comprising: an apparatus configured to obtain a scalp electroencephalogram (EEG) data set from the subject; and, a controller operably connected to the apparatus, which controller comprises a processor, and a memory communicatively directly or remotely coupled to the processor, the memory storing non-transitory computer executable instructions which, when executed on the processor, perform operations comprising: generating at least one dynamic brain network model (DNM) for the subject using the scalp EEG data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

11 . The system of claim 10, wherein a wearable device comprises the apparatus.

12. The system of claim 10, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject.

13. The system of claim 10, wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

14. The system of claim 10, wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

15. The system of claim 10, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject.

16. The system of claim 15, wherein the report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease.

17. The system of claim 15, wherein the report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.

18. A computer readable media, comprising non-transitory computer executable instructions which, when executed by a processor, perform operations comprising: generating at least one dynamic brain network model (DNM) for a subject using a scalp electroencephalogram (EEG) data set obtained from the subject; determining at least one sink index (SI) for the subject using the DNM; and determining whether the subject has frontotemporal dementia (FTD), Alzheimer’s disease, or does not have dementia using one or more SI ratios of brain regions associated with FTD to other brain regions.

19. The computer readable media of claim 18, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: generating the DNM for the subject using a scalp electroencephalogram (EEG) data set obtained from the subject.

20. The computer readable media of claim 18, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: determining an SI heatmap for the subject.

21 . The computer readable media of claim 18, wherein the SI for the subject comprises a quantitative measure of each brain region in a brain of the subject computed from imaging data obtained from the subject.

22. The computer readable media of claim 18, wherein the SI ratios comprise a ratio of FT nodes to Central, Parietal, and Occipital (CPO) nodes.

23. The computer readable media of claim 18, wherein the non-transitory computer executable instructions which, when executed by the processor, further perform operations comprising: outputting a report indicating the neurological disorder status of the subject.

24. The computer readable media of claim 23, wherein the report comprises a recommendation to administer at least one therapy to the subject when the neurological disorder status indicates that the subject has FTD and/or Alzheimer’s disease.

25. The computer readable media of claim 23, wherein the report comprises a recommendation to discontinue administering at least one therapy to the subject when the neurological disorder status indicates that the subject does not have dementia.