Dissecting unsupervised learning through hidden Markov modeling in electrophysiological data.

"data will be made available upon reasonable request 10 6084/m9 figshare 23283656 supplemental figs s1-s5: https://doi org/10 6084/m9 figshare 23283656 . all the codes used for the generation of synthetic data and for the analysis of both synthetic and real data are available at https://github com/lauramasaracchia/hmm_explore ."

"both are implemented in the hmm-mar toolbox publicly available on github 1 in our analyses we manipulated: the respective model hyperparameters: the order p for the hmm-mar and the lags structure for the hmm-tde defined by the width l and the inter lags steps s (see below for definitions).; all the codes used for the generation of synthetic data and for the analysis of both synthetic and real data are available at https://github com/lauramasaracchia/hmm_explore ."

"DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the authors."

"EC | European Research Council (ERC)10.13039/501100000781; NIHR | NIHR Oxford Biomedical Research Centre (OxBRC)10.13039/501100013373; Novo Nordisk Fonden (NNF)10.13039/501100009708; Wellcome Trust (WT)10.13039/100010269"