in house developed neurophysiological biomarker toolbox nbt  (MathWorks Inc)

Journal: bioRxiv

Article Title: E/I ratio and net E+I strength are differentially affected across brain disorders

doi: 10.1101/2025.08.15.670484

Figure Lengend Snippet: Both seizure zone ( A , B , E–H ) and non-seizure zone ( C , D , I-L ) electrodes show increases in E/I HLP and E+I HLS during ictal periods. Biomarker values were averaged across seizure (SOZ)/non-seizure zone (nSOZ) electrodes, and a paired t -test was run across subject-level averages. Circles ( A–D ) correspond to frequency bins where comparisons stayed significant after FDR correction ( q =0.05). ( E–L ) A representative subject was selected, and sample traces of the oscillatory power time series were plotted for a selected frequency band (10.5-13.4 Hz), along with the bimodal power distributions, to illustrate the changes in oscillatory dynamics between the ictal and interictal states.

Article Snippet: EEG analyses were initially pre-processed using the in-house developed Neurophysiological Biomarker Toolbox (NBT) written in MATLAB.

Techniques: Biomarker Discovery

Journal: bioRxiv

Article Title: E/I ratio and net E+I strength are differentially affected across brain disorders

doi: 10.1101/2025.08.15.670484

Figure Lengend Snippet: (A) During ictal periods, the seizure zone (SOZ) exhibits significantly higher E/I HLP compared to non-seizure zone (nSOZ). (C) During interictal periods, on the other hand, the seizure zone has lower E/I HLP . (B , D) E+I HLS is significantly higher in the seizure zone regardless of the seizure status. Biomarker values were averaged across seizure (SOZ)/non-seizure zone (nSOZ) electrodes, and a paired t -test was run across subject-level averages. Circles ( A–D ) correspond to frequency bins where comparisons stayed significant after FDR correction ( q = 0.05). ( E–L ) A representative subject was selected, and sample traces of the oscillatory power time series were plotted for a selected frequency band (10.5-13.4 Hz), along with the bimodal power distributions, to illustrate differences in oscillatory dynamics between SOZ and nSOZ.

Article Snippet: EEG analyses were initially pre-processed using the in-house developed Neurophysiological Biomarker Toolbox (NBT) written in MATLAB.

Techniques: Biomarker Discovery

Journal: bioRxiv

Article Title: E/I ratio and net E+I strength are differentially affected across brain disorders

doi: 10.1101/2025.08.15.670484

Figure Lengend Snippet: (A) Logistic regression models reveal that patients with Alzheimer’s, ADHD, or ASD can be distinguished from healthy controls by alterations in E/I HLP and E+I HLS . Significance for each disorder was assessed by comparing the average AUC-ROC (black dot) across 50 cross-validation folds against a null distribution generated with 500 random permutations (gray violin plot). Black asterisk reflects significance after FDR ( q = 0.05) correction. (B) When training logistic regression separately for E/I HLP and E+I HLS , we found Alzheimer’s disease showing mostly disruptions in E/I ratio, while ASD and ADHD more prominent changes in E+I density. To illustrate how the relative contribution of E/I HLP and E+I HLS differs across the three disorders, we plotted disorder-specific case-control averages of each biomarker across the frequency spectrum ( C–H ). Asterisks above a trace mark frequency bins in which patient and control means differ significantly (two-sample t -test, FDR-corrected at q = 0.05, performed separately for every disorder-by-biomarker combination).

Article Snippet: EEG analyses were initially pre-processed using the in-house developed Neurophysiological Biomarker Toolbox (NBT) written in MATLAB.

Techniques: Biomarker Discovery, Generated, Control