Albrecht2019: Schizophrenia Classification Dataset with Modified Simon Task

The Albrecht2019 dataset originates from a study investigating reinforcement learning under cognitive conflict in individuals with schizophrenia (PSZ) and healthy controls. The dataset includes both behavioral and EEG recordings collected during a modified Simon task, which introduces response conflict as an implicit cognitive cost during reinforcement learning.

A total of 78 participants took part in the study: 46 individuals with a DSM-IV diagnosis of schizophrenia or schizoaffective disorder, and 32 healthy controls. EEG data were recorded using a 64-channel Brain Vision system at a 1000 Hz sampling rate. Data were preprocessed and artifact-corrected using the EEGLAB pipeline, and ICA was applied for eye-blink removal. EEG epochs were extracted around stimulus and feedback events to capture conflict-evoked and prediction-error-related activity, particularly in the theta band (4–7 Hz).

Participants completed a reinforcement learning version of the Simon task. On each trial, a stimulus was associated with probabilistic reward or punishment outcomes, modulated by whether the trial involved a congruent or conflict-inducing response. This design enabled the dissociation of positive and negative prediction error (PE) learning biases under cognitive effort. A subsequent transfer phase assessed stimulus preferences to infer learning outcomes.

Paper

Albrecht, M. A., Waltz, J. A., Cavanagh, J. F., Frank, M. J., & Gold, J. M. (2019). Increased conflict-induced slowing, but no differences in conflict-induced positive or negative prediction error learning in patients with schizophrenia. Neuropsychologia, 123, 131-140.

DISCLAIMER: We (DISCO) are NOT the owners or creators of this dataset, but we merely uploaded it here, to support our's (EEG-Bench) and other's work on EEG benchmarking.

Dataset Structure

• data/ contains the experiment data.
• DataDictionary.txt explains the meaning of columns and codes in various files.
• DeID_Dems.csv gives an overview of the subjects' demographic attributes.
• MMN_ElectrodeCoords_MA2.ced contains information about the EEG electrode positions.
• Pre-processing_Pseudo.txt explains the pre-processing pipeline in pseudocode applied on the EEG data.
• scripts/ contains the MATLAB scripts that were used to execute the experiment.

Filename Format

CC_EEG_s[PID]_[CONDITION].mat

PID is the patient ID (e.g. 101), while CONDITION denotes the condition (P means "with schizophrenia", N means "without schizophrenia"). All patients with PID <= 147 have schizophrenia and all patients with PID >= 149 do NOT have schizophrenia and hence belong to the control group.

Fields in each File

A .mat file can be read in python as follows:

from scipy.io import loadmat
filename = "CC_EEG_s101_P.mat"
mat = loadmat(filename, simplify_cells=True)

(A field "fieldname" can be read from mat as mat["fieldname"].)

Then mat contains (among others) the following fields and subfields

• EEG
  • data: EEG data of shape (#channels, time_len). E.g. a recording of 30 minutes with 64 channels and a sampling rate of 1000 Hz will have shape (64, 1800000).
  • event: Contains a list of dictionaries, each entry (each event) having the following description:
    • latency: The onset of the event, measured as the index in the time-dimension time_len. As noted in the paper, each stimulus has a response window of 850 ms.
    • type: The type of event. This type will be a number of up to 3 digits. In order to decode them, please refer to the Triggers section in DataDictionary.txt. Typically, a trial starts with a Stimulus Trigger, followed by a Response Trigger (or 105 if no response was given, i.e. the patient did not press a button) and then a Feedback Trigger.
  • srate: Sampling Rate (Hz)

Alternatively, event information for any CC_EEG_s[PID]_[CONDITION].mat file can also be read from the accompanying CC_Beh_s[PID]_[CONDITION].mat file. There, trials are organized in the 4 blocks described in the paper. Information about a trial can be read as

beh_mat = loadmat("CC_Beh_s[PID]_[CONDITION].mat", simplify_cells=True)
block = 0          # can be 0, 1, 2 or 3
trial_number = 5   # number of the trial
trial_info = beh_mat["Beh"][block]["data_block"][trial_number]

trial_info can then be decoded using the Behavioral Data section of DataDictionary.txt.

License

By the original authors of this work, this work has been licensed under the PDDL v1.0 license (see LICENSE.txt).