https://saheedfaremi.com Saheed Faremi Notes on EEG microstate research, deep learning methods, and engineering practice. 2026-06-20 https://saheedfaremi.com/blog/questions-behind-vae-eeg-microstates Learning EEG microstates with a variational autoencoder: representation, the explained-variance gap, and the locus of recurrence 2026-06-20 A methods note on a Conv-VaDE pipeline for EEG microstates: why topographic images, why the model trails modified k-means on explained variance, the invariance of sequence statistics across deep variants, and the distinction between recurring patterns and recurrent dynamics. A methods note on a Conv-VaDE pipeline for EEG microstates: why topographic images, why the model trails modified k-means on explained variance, the invariance of sequence statistics across deep variants, and the distinction between recurring patterns and recurrent dynamics.

]]> https://saheedfaremi.com/blog/eeg-preprocessing-stages EEG preprocessing, stage by stage 2026-05-15 Walking through the seven preprocessing stages that take a raw LEMON recording from disk to the input my Conv-VaDE actually sees. Walking through the seven preprocessing stages that take a raw LEMON recording from disk to the input my Conv-VaDE actually sees.

]]> https://saheedfaremi.com/blog/four-quadrant-evaluation Four ways to score the same model 2026-05-15 Cluster-validity scores disagree on the same model. So do the spaces you evaluate them in. The four-quadrant framework keeps both axes honest. Cluster-validity scores disagree on the same model. So do the spaces you evaluate them in. The four-quadrant framework keeps both axes honest.

]]> https://saheedfaremi.com/blog/topomap-masking-convention The four corners of every topomap are not brain signal 2026-05-15 Every 40x40 topomap in the pipeline has a constant-zero ring of padded pixels outside the inscribed head-disc. Every loss and metric in the project masks them out. Here is why. Every 40x40 topomap in the pipeline has a constant-zero ring of padded pixels outside the inscribed head-disc. Every loss and metric in the project masks them out. Here is why.

]]>