Date created: | Last Updated:

Category: Project

Description: Introduction: ========== EEG data that was collected from the Tata Longitudinal Study of Aging (TLSA) cohort from urban communities in Bangalore between July 2016 and July 2019. In addition to the TLSA cohort, two other datasets were used in one study (Murty et al., 2020, Neuroimage). This consisted of EEG data from young subjects (mainly students of IISc, used in Murty et al., 2018, JNeurosci) and another cohort of subjects less than 49 years old. These are called “VisualGamma” and “AgeProjectRound1” projects. These databases are not maintained. EEG data were collected in 350 sessions from 279 unique subjects. For each session, the subjects were clinically diagnosed by psychiatrists (authors BN/AML in Murty et al., 2020, Neuroimage) and/or a neurologist (author MJ) as cognitively healthy, MCI or AD through clinical history and a semi-structured clinical interview. Gamma Protocol (also called SF_ORI since spatial frequency and orientation were varied) 1. Full Dataset: 350 sessions (279 unique subjects) 2. Discarded: 40 sessions (EEG was not even analysed for these sessions) a. Did not complete the experiment: 4 b. Removed from analysis due to recording errors, poor vision etc: 17 c. Could be replaced with a cleaner followup/baseline: 6 d. Label (HV/MCI/AD) pending or had a discrepancy: 9 e. Age less than 50 years: 1 f. Assigned as a repeat even though the baseline was later discarded: 3 Remaining good sessions: 310 (257 unique subjects, HV/MCI/AD: 236/15/6) 3. No useful protocols: 11 (data was analysed but after removal of bad trials/electrodes, the session could not be used) Remaining good sessions: 299 (247 unique subjects, HV/MCI/AD: 227/14/6) 4. Subjects with repeats: 299-247 = 52. a. 3 subjects had different labels in the two sessions b. 1 subject was MCI on both sessions Remaining: 48 subjects who were healthy in both baseline and follow up. Data format ============ 1. rawData (220 GB): files originally generated by EEG data acquisition system. All data is extracted from here. Not available in this folder. 2. SegmentedData (272 GB): Segments of data around the stimulus onset are extracted and saved from rawData. Not available in this folder. 3. cleanData (153 GB): Bad trials are removed from segmentedData using the pipeline described in Murty et al., 2020, Neuroimage. Not available in this folder. 4. decimatedData (43.8 GB): EEG data in cleanData was decimated by a factor of 10 and then saved. Available in TLSAEEGProject/decimatedData [this decimated data is provided in the present repository] 5. analyzedData: Intermediate data kept in each Project Folder.