This repository contains scripts for the paper:
"Closed-loop auditory stimulation targeting alpha and theta oscillations during REM sleep induces phase-dependent power and frequency changes"
by Valeria Jaramillo, Henry Hebron, Sara Wong, Giuseppe Atzori, Ullrich Bartsch, Derk-Jan Dijk*, Ines R. Violante* (* contributed equally).
Preprint can be found here: biorxiv DOI
Raw data and data to create the figures can be found here: Zenodo DOI
Required toolboxes/functions to run the scripts:
- eeglab v2021.1
- fieldtrip revision ebc13229d
- circular statistics
- eBOSC
- linspecer
- shadedErrorBar
- echt
- DataViz
- colorGradient
- distinguishable colors
- ScientificColourMaps
A1_writeedf_for_scoring: write .edf file that is used for sleep scoring (re-referencing 8 EEG channels used for sleep scoring to mastoids), you need to export .eeg file as .edf file using BrainVisionRecorder before running this script (.edf file is needed)
A2_extract_triggers: extract AEP and CLAS triggers from .eeg file, save in _nm.mat file
B1_load_filter_data: load and filter sleep EEG data from .eeg file, save as _fil.fdt and .set file
C1_artcorr_extract_REM: load filtered data, sleep scoring, and phasic/tonic scoring data, visualize REM sleep EEG using eeglab, visually identify and interpolate bad channels, visually mark segments that contain artefacts, save _goodREM.fdt and .set files that only contain 'good' (artefact-free) REM samples, save _goodREM.mat file containing information about sleep and phasic/tonic scoring, bad channels, and artefacts
D_ICA: perform ICA on _goodREM data using eeglab using AMICA1.6 plugin
E_manual_ICA: visualize independent components using eeglab, visually identify eye movement, cardiac, muscle activity, and channel noise components and remove those components, save as _manual_ICA.fdt and .set files
F_insert_REM_after_ICA_rereferencing: put back cleaned REM sleep data (after ICA) into the whole-night data and re-reference to the average across all channels
G_goodtrigs: save triggers that occurred during a 'good' REM sample in _nm_good.mat
H_ffttot_sleep: calculate power spectral density for each sleep epoch
A_load_filter_data_wake: load and filter wake EEG data from .eeg file (raw data, provided), save as _fil.fdt and .set file
B_extract_triggers_wake: extract AEP triggers from .eeg file (raw data, provided), save in _nm.mat file
C_artcorr_wake: load filtered data, visualize EEG using eeglab, visually identify and interpolate bad channels, visually mark segments that contain artefacts, save _goodwake.fdt and .set files that only contain 'good' (artefact-free) wake samples, save _goodwake.mat file containing information about bad channels, and artefacts
D_ICA_wake: perform ICA on _goodwake data using eeglab using AMICA1.6 plugin
E_manual_ICA: visualize independent components using eeglab, visually identify eye movement, cardiac, muscle activity, and channel noise components and remove those components, save as _manual_ICA.fdt and .set files
F_insert_wake_after_ICA_rereferencing: re-reference wake data to the average across all channels
G_goodtrigs: save triggers that occurred during a 'good' wake sample in _nm_good.mat
H_ffttot_wake: calculate power spectral density for each wake epoch
make analyses, run after preprocessing, run psd analyses first, rest can be run in any order
perform AEP analyses for REM and wake data, run in order A-B
A_ERP_sub_REM: extract AEP REM data for each participant
A_ERP_sub_wake: extract AEP wake data for each participant
B_ERP_allsub_REM: save AEP REM data for all participants
B_ERP_allsub_wake: save AEP wake data for all participants
perform connectivity analyses, run in order A-D
A_phase_connectivity: calculate the phase using Hilbert transform for trial data
B_PLV_sub: calculate PLV and PLI for all channel pairs for each participant
C_PLV_allsub: save PLV and PLI for all participants
D_cluster_analysis_SnPM_connectivity: calculate lme for each electrode with cluster correction for alpha CLAS
D_cluster_analysis_SnPM_connectivity_theta: calculate lme for each electrode with cluster correction for theta CLAS
calculate ISI's to deterime stimulation frequency, detect oscillations to determine individual peak frequency, run in order A-C
A_ISI_allsub_allch: calculate ISI's and save for all participants
B_eBOSC: detect oscillations for each epoch
C_eBOSC: extract oscillations with at least 3 cycles and 300 ms duration, plot histogram and detect peak
perform frequency analyses, run in order A-C
A_instantaneous_frequency: calculate instantaneous frequency for REM and wake data for each participant
B_freq_allsub: calculate frequency for ON and OFF windows and save for all participants
C_cluster_analysis_SnPM_frequency: calculate lme for each electrode with cluster correction for alpha CLAS
C_cluster_analysis_SnPM_frequency_theta: calculate lme for each electrode with cluster correction for theta CLAS
calculate number of trials, run in order A-B
A_ntrials_ERPs: save number of AEP trials (all, phasic, tonic, thirds of night) for all participants
B_ntrials_nm: save number of CLAS trials (all, phasic, tonic, thirds of night) for all participants
perform power analyses, run in order A-C
A1_Power_rem_wake: average power across all, phasic, and tonic epochs, and all wake, EO, and EC and save for all participants
A2_Power_ON_OFF_goodREM: perform power analysis for all trials of each condition and for each participant
B_Power_ON_OFF_allsub_allch: average across all good trials, all good phasic and all good tonic trials and save for all participants
C_cluster_analysis_SnPM_power: calculate lme for each electrode with cluster correction for alpha CLAS
C_cluster_analysis_SnPM_power_theta: calculate lme for each electrode with cluster correction for theta CLAS
perform phase-locking accuracy analyses, run in order A-B
A_Phase_sub_allch: perform phase analysis for all triggers of each condition and for each participant
B_resultant_allsub_allch: calculate resultant and mean phase for each participant, condition and channel and save for all participants
calculate sleep stage and phasic/tonic percentages, run in order A-B
A_Table1_Sleep_parameters: reads in .txt file containing sleep scoring labels, calculate classical sleep parameters B_Table1_phasic_tonic_conditions: calculate percentage of phasic and tonic REM sleep for each participant and for ON and OFF windows for each condition, save for all participants
make figures, run after analyses
Figure1C_phasic_tonic_epoch: plot example 10-s data for phasic and tonic REM sleep
Figure1D_Suppl_Figure1_ntrials_across_night_erps: make boxplots for number of AEP and CLAS trials (all and phasic)
Figure2_Suppl_Figure2_phasic_tonic_psd_AEPs: plot power spectrum and AEPs for phasic, tonic and wake EO (eyes open) and EC (eyes closed)
Figure3A_3F_ind_psd: plot power spectrum for each individual in alpha and theta range
Figure3B_3G_eBOSC_frequency_histogram: plot histogram of detected oscillations and stimulation frequencies for example participant
Figure3C_3H_eBOSC_frequency_corr_nwaves: plot correlation scatter plot between individual oscillation peak frequency and peak stimulation frequency
Figure3D_3I_polarhistogram: plot polar histogram of mean phase and resultant across participants
Figure3E_3J_resultant_topo: plot resultant for each electrode across scalp, test non-uniformity across circle
Figure4A_4G_Suppl_Figure5-9and12-17_Power_ON_OFF_allsub_topo_boxplots: plot topography of power change lme F-values and significant electrode clusters
Figure4B_4C_Power_ON_OFF_change_psd_phasictonic_alpha: plot power changes across spectrum for alpha CLAS
Figure4D_4J_Suppl_Figure10and18_Frequency_ON_OFF_allsub_topo: plot topography of frequency change lme F-values and significant electrode clusters
Figure4E-F_4K-L_Frequency_ON_OFF_change_boxplots_time: make boxplots for frequency change and plot frequency change over time across different conditions
Figure4H_4I_Power_ON_OFF_change_psd_phasictonic_theta: plot power changes across spectrum for theta CLAS
Suppl_Figure3_ISI_autocorr: plot autocorrelation across ISI's
Suppl_Figure11_19_connectivity: plot topography of connectivity change lme F-values and significant electrode clusters, plot significant connections and boxplots for different conditions
Suppl_Table1_questionnaires: make table for KSS and VAMS changes from evening to morning