Sitemap

The page you were looking for doesn’t exist.

Applied machine learning, time-series classification, distribution shift, and reproducible research software, with EEG as the primary research domain.

A working note on using regime-switching linear state-space models for non-invasive brain-signal decoding.

A direct guide to reading when EEG domain adaptation is useful and when it causes negative transfer.

A practical note on Wasserstein, MMD, and Energy Distance for quantifying session-to-session shift in EEG pipelines.

A practical view of how multiple source sessions can help or hurt cross-session EEG transfer.

A concise bridge from classical ICA to multilinear/tensor ICA design choices.

A plain comparison of four domain-adaptation methods through the object each method tries to align.

A compact map of the main ICA objective functions and their algorithmic implications.

A simple explanation of why EEG covariance matrices are useful features and how tangent-space mapping makes them usable by standard classifiers.

A direct walkthrough of how CSP and filter-bank CSP turn motor-imagery EEG trials into classifier features.

Efficient whitening using Cholesky decomposition in the eegwhiten package.

A plain guide to fitting EEG feature extractors without leaking test information.

A deep dive into the PCA whitening algorithm implemented in the eegwhiten package.

Understanding Zero-phase Component Analysis (ZCA) whitening in eegwhiten.

Using Singular Value Decomposition (SVD) for robust whitening in eegwhiten.

R Package Developer
Whitening and covariance-alignment tools for EEG features, with reusable fit/transform models, shrinkage, robust covariance estimators, Euclidean alignment, and tangent-space utilities.

R Package Developer
A unified R framework for EEG/BCI domain adaptation, including TCA, SA, CORAL, GFK, MIDA, Riemannian transport, optimal transport, and shift metrics.

R Package Developer
A train/test-safe EEG feature extraction toolkit for BCI pipelines, including CSP/FBCSP, tangent-space covariance features, bandpower, Hjorth, ATM, MVAR, and MSVAR features.

R Package Developer
A physics-constrained EEG simulator and covariance-aware augmentation audit package for cross-session BCI research.

Rank: 98/2767 (Top 4%) | Silver Medal
Developed a deep learning pipeline using EfficientNet and Weighted Ensembling to classify seizures and harmful brain patterns from EEG signals.

Python Package Developer
Python implementation of DA4BCI, providing EEG/BCI domain adaptation methods, shift metrics, SPD geometry tools, Page-Hinkley drift detection, and plotting utilities.

Python Package Developer
Python port of TensorEEG for physics-constrained EEG simulation, SPD covariance augmentation, fidelity auditing, and manifest replay.

Python Package Developer
Reproducible comparison of pooling, weighting, and source-selection strategies for transfer across EEG sessions.

Dashboard Developer
A Streamlit dashboard for comparing multi-source domain adaptation strategies in cross-session EEG classification.

Dashboard Developer
An offline analysis dashboard separating fixed-model degradation, domain-adaptation benefit, and the remaining gain from retraining on later EEG sessions.

Application Developer
A sample-paced pseudo-online replay application for motor-imagery EEG, with CSP, FBCSP, tangent-space pipelines, Euclidean Alignment, and progressive performance visualization.

Application Developer
A dependency-free browser visualizer for understanding how domain adaptation methods move source and target distributions.

Interactive electrode-layout design with bounded optimization, forward simulation, and model-based evaluation.

An experimental macOS application integrating local person tracking, camera capture, and Bluetooth gimbal control.

Manuscript in preparation
A tensor-based statistical framework extending MCCA to high-dimensional datasets, preserving structural information in multi-view neuroimaging analysis.

Manuscript under review
Proposing a ‘Linear-First’ decision rule using Paired Non-Inferiority Tests (TOST) to balance decoding accuracy against computational cost.

Manuscript under review
A geometric framework that diagnoses BCI performance degradation by separating signal drift into raw sensor variability and feature-space distortions.

Manuscript under review (solo author)
A cross-participant study of MI-EEG system selection: single decoders, uncertainty-routed decoder pairs, and ensembles compared by final accuracy and execution cost, with calibration-based selection tested on held-out participants.

Published in Frontiers in Human Neuroscience, 28 Aug 2026
A matched comparison of global and selective source-session strategies for within-subject, cross-session motor-imagery EEG transfer.

Ph.D. Dissertation, University of Massachusetts Boston (Aug 2026)
Reliable MI-EEG decoding as a sequence of representation, transfer, drift measurement, adaptation, and recalibration choices.

Authors: Yiming Shen, David Degras (University of Massachusetts Boston)

Authors: Yiming Shen, David Degras (University of Massachusetts Boston)

Presenter: Yiming Shen
Poster number: G5-Tues