sleep scoring based on intracranial recordings (in collaboration with Yuval Nir and Hanna Hayat)
GENERAL DESCRIPTION - In sleep sessions on which no PSG was recorded (or recorded signals were noisy), conventional scalp EEG-based sleep staging was made impossible. Instead, this toolbox allows sleep scoring by using depth EEG. We calculated spectrograms with a 30 sec window (no overlap) from 0 to 40Hz and averaged the power in the delta band (0-4Hz).
FINER DETAILS - sleepScore_example.m walks you through the steps of iEEG based sleep scoring.
Step 1 generate a hypnogram for all available channels - using plotHypnogram_perChannel() (create_sleepHypnogram_per_pt() is my wrapper function for that purpose).
Step 2 select the best channels to perform sleep-scoring with. The best candidates would be frontal, cingulate, parietal as medial as possible. Look for a contact with prominent spindle power. Update this in an XLS table that will be fed into sleepScoring_iEEG_wrapper.m
Step 3 sleepScoring_iEEG_wrapper uses evaluateDelta() to perform sleep scoring on the selected channels and produces figures that can be used to judge how well it performed. Next - use manualValidationSleepScoring() to fine tune the scoring (remove/add epochs).
Finally - this code will generate a figure with the power statistics of NREM/REM sleep based on this channel.