Decoding rapidly presented visual stimuli from prefrontal ensembles without report nor post-perceptual processing

The following is the dataset used in the study "Decoding rapidly presented visual stimuli from prefrontal ensembles without report nor post-perceptual processing." published in Neuroscience of consciousness in 2022 (https://doi.org/10.1093/nc/niac005) This dataset features multiunit activity (MUA) recordings from two adult male macaque monkeys, captured via Utah arrays implanted in specific brain regions: the ventrolateral prefrontal cortex (vlPFC, area 45a) and the parietal cortex (PPC, areas 7a/7b). The recordings were made as the monkeys engaged in a passive visual task, where they were required to maintain their gaze on a central fixation point while a sequence of images was displayed. These images were presented in two distinct formats: a rapid serial visual presentation (RSVP) with a stimulus onset asynchrony (SOA) of 100 ms, and a slower-paced presentation with SOAs of 400 ms for Monkey H and over 900 ms for Monkey A. The primary aim of collecting this data was to investigate how the vlPFC encodes visual stimuli under two different conditions: one that allows time for cognitive reflection on each image (slow-paced) and another that likely precludes such reflection due to rapid presentation (RSVP). Additionally, PPC data was gathered to compare the decoding signal strength across an equivalent number of recording channels. For Monkey A, PPC and vlPFC data were collected concurrently, whereas for Monkey H, the PPC data was obtained in separate sessions. ________________________________________________________________________________!!! IMPORTANT !!! For users seeking a more navigable dataset, click here to access this Python notebook. This notebook provides tools for renaming variables and cleaning unused columns, allowing for a more tailored data analysis experience without altering the primary dataset files. This notebook is designed to enhance your data analysis experience by providing tools for renaming variables and cleaning unused columns without altering the original dataset files. Generated Dataframes:cleaned_monkeyA.pkl: Contains simultaneous recordings from the Prefrontal Cortex (PFC) and Posterior Parietal Cortex (PPC) for Monkey A. cleaned_monkeyH_PFC.pkl and cleaned_monkeyH_PPC.pkl: Contain recordings from Monkey H's PFC and PPC, respectively, collected in different sessions.Dataset Structure in Cleaned Files:Each line in the cleaned datasets - InterStimulusInterval: Measures the time in seconds between the onset of one stimulus and the onset of the next. - TrialIndex: An integer that uniquely identifies each trial. All stimuli presented within the same trial share this index, facilitating the grouping of data by trial. - RecordingDayIndex: An integer that uniquely identifies each recording day, allowing analyses to be segmented by specific experimental sessions. - StimulusIdentity: A number between 0 and 17 that identifies the specific image displayed during a trial. This identifier is used to correlate specific stimuli with behavioral and neural responses. - StimulusPosition: An ordinal number indicating the position of the stimulus within a trial sequence. The first stimulus in a sequence is numbered 1, which is important for studying response patterns to stimulus order. - StimulusDuration: Time in seconds that the stimulus is physically present on the screen, as measured by the photodiode. - SpikeTimes_vlPFC: Lists the times of neural spikes in seconds, relative to the stimulus occurrence for each of the 96 channels of the vlFPC array. This detailed recording from the Ventrolateral Frontal Parietal Cortex provides insight into neural activity during stimulus presentation. - SpikeTimes_PPC: Similar to `SpikeTimes_vlFPC`, but for the Parietal array. This records the spike times in seconds relative to the stimulus occurrence, enabling comparative studies of neural dynamics across different brain regions. Each entry in the **SpikeTimes...** columns is a list of 96 lists. Each of these 96 lists contains the times at which spikes were detected relative to the stimulus onset. The time interval considered for detecting a spike is between -0.1 seconds and 0.7 seconds relative to stimulus onset. ________________________________________________________________________________Original Dataset Variables Description The following section provides a detailed description of the original variables in the RSVP dataset. The dataframes monkeyA.pkl, monkeyH.pkl and monkeyH_PPC.pkl contain all the necessary information from which the other files are computed. They consist of tables of stimuli ordered by their chronological appearance. Each line of a dataframe corresponds to one stimulus that is either in isolation or part of a sequence. The .pkl dataframes contains the following entries: -Spikes is the time of the spikes relative to the time of the stimulus occurring in each channel of the VLFPC array (in seconds). It is a list corresponding to the 96 channels of the Utah array.- PPC_Spikes is the same as Spikes but for the parietal array- TrialID is the time in milliseconds separating a stimulus onset from the next one (as measured by the photodiode). Some trials in monkeyH.pkl have an inter-stimulus interval of 200 ms. These trials where not analyzed in the original study.- ItemID is a number indicating the ordinal number of the current stimulus in a trial. The first stimulus in a sequence has the number 1- StimID is a number between 0 and 17 indicating the identity of the image displayed- TrialIDcount gives the same integer for all stimuli presented in a given trial- sesID gives the same integer for all stimuli presented in a given recording day The other files are precomputed firing rates, classifier coefficients, and predictive probabilities that are used in the study. They can be recomputed from the .pkl files using the code reproducing the analysis of the paper: https://github.com/jobellet/fast_and_rich_decoding_in_VLPFC