DOI: 10.5061/dryad.cjsxksnfh

Description:

This dataset includes post-processed data derived from the raw data using Suite2p, original MATLAB files obtained during the BMI experiments, and pandas data frames containing analysis results. These results are the basis for the figures presented in the paper.

Data and File Structure:

The dataset comprises the following types of files:

1. ZIP files for each session and each mouse: The filenames include the mouse identifier (e.g. “m13”), date (e.g. “221112_D01”), and a suffix indicating whether the file contains post-processed imaging data (suffix “_suite2p”) or original BMI execution data (suffix “_bmi”).
   • Example filenames: m13_221112_D01_suite2p.zip, m13_221112_D01_bmi.zip
2. CSV files: These files contain specific analysis results that are not stored in Parquet format. CSV files store tabular data in plain text, where each line corresponds to a data record, and each record consists of fields separated by commas.
3. Parquet files: These files include dataframes with analysis results utilized in the paper’s figures. Parquet is a columnar storage file format optimized for use with data processing frameworks like pandas.
4. npy files: These files are used for storing large, multi-dimensional arrays of numerical data, typically used in scientific computing with Python.
5. MAT files: These files store raw data collected during BMI experiments in MATLAB format, allowing for easy use and further analysis in MATLAB.

Content of the Suite2p ZIP Files:

• bad_frames_dict.npy: Contains information (in form of a python dictionary) about frames that were deemed bad during processing. It contains the index of the frames that are “bad”: ‘bad_frames_index’ (array with length T=time), and a boolean array (size T) indicating “1/True” on indeces of the frames that were “bad”. ‘bad_frames_bool’.
• bad_frames.npy: Array size T, that lists bad frames identified during the imaging sessions. Index of the bad frames to be used during registration on neurons.
• dff.npy: Array with NxT (N=neurons, T=time) that contains delta F over F values, which represent changes in fluorescence intensity.
• direct_neurons.npy: Dictionary that identifies neurons directly controlling the BMI. Contains ’E1’: Index of the neurons (N) of the ensemble E1, ‘E2’: Index of the neurons (N) of the ensemble E2, ‘exclude’: Index of the neurons that must be excluded (redundant or merged with others), ‘added_neurons’: and index of the neurons that were added posthoc by the experimenter}
• F.npy: NxT array that contains raw fluorescence data obtained by the Suite2p.
• Fneu.npy: NxT array that contains neuropil fluorescence data, which helps in subtracting background fluorescence, obtained by the Suite2p.
• iscell.npy: Array with Nx2 that identifies cell ROIs (Regions of Interest) as neurons if first column is a 1, and gives the measure of “being a neuron” in the second column.
• iscell_old.npy: Array with Nx2 that contains the original classification of cell ROIs done by Suite2p
• ops.npy: Contains the final operational parameters used in the Suite2p processing.
• ops_before_1st.npy: Operational parameters used in the first iteration of Suite2p processing.
• spks.npy: NxT array that contains inferred spike rates of neurons obtained with suite2p.
• spks_dff.npy: NxT array  that contains inferred spike rates derived from delta F over F values (dff.npy).
• stat.npy: N array that contains info about identified ROIs as obtained by Suite2p.
• stim_time_dict.npy: Dictionary that contains the array (T) with indexes of stims:”stim_index” and the bool array (T) that indicate with a “1” when the stim happened (“stim_bool”).
• targettimedict.npy: Dictionary that contains the array (T) with indexes of when the target was achieved:”target_index” and the bool array (T) that indicate with a “1” when the target was achieved (“target_bool”).

Content of the BMI ZIP Files:

• folder motor: Contains behavioral motor data. It contains 2 files with behavioral data. The filenames include the date (e.g. “221112_D01”), a suffix indicating whether the file contains the “XY” positions of the animal or the metadata “Trigger”, the mouse identifier (e.g. “13”), and a  suffix indicating if the behavior correspond to the “Baseline” (calibration) or the “BMI” (e.g. “221112_XY_13_BMI”).
• BaselineOnline.mat: Array with NxT raw data of all the online neurons captured during calibration.
• BMI_online.mat: Data from the BMI online session. Contains “bData” and “data”. 
  • bData: Metadata to run the BMI based on BaselineOnline.mat data
  • data: 
    • data.cursor, T array with the BMI neural cursor
    • data.fb_freq, T array with the frequency of the auditory feedback (mapped to the cursor)
    • data.bmiAct, NxT neural activity of the direct neurons.
    • data.baseVector, T baseline vector obtained and used during the BMI
    • data.selfHits, T boolean array to indicate with a “1” that a target has been acquired
    • data.selfDRstim, T boolean array to indicate with a “1” that a stim has been delivered (through BMI)
    • data.vector_stim, Array of indeces where to delivered a predetermined stim (obtained by randomly assigning indeces)
    • data.vectorWater, T boolean array to indicate with a “1” that water reward has been delivered. There was none in these experiments.
    • data.randomDRstim, T boolean array to indicate with a “1” that a stim has been delivered (through “vector_stim”)
    • data.trialStart,  T boolean array to indicate with a “1” that a trial starts.
    • data.timeVector,  T boolean array to indicate with time lapse from the previous index.
    • data.selfTargetCounter, Counter of targets achieved with BMI
    • data.selfTarget_DR_stim_Counter, Counter of stims delivered with BMI
    • data.sched_random_stim, Counter of stims delivered from “vector stim”
    • data.Water_Counter, Counter of water rewards
    • data.trialCounter, data.frame, Counter of frames
• BMI_target_info.mat: Metadata with the Target information from the BMI session (redundancy with bData)
• roi_data.mat: Structure containing the spatial distribution of neurons.
  • im_sc_struct, structure with info about the resizing of the background image. Metadata.
  • plot_images, Green and Red channel reference image
  • roi_data: Calibration roi-masks for all the neurons recorded
    • roi_data.im_bg, background image (reference green image)
    • roi_data.num_rows, roi_data.num_cols, number of pixels in the image
    • roi_data.num_rois, number of rois recorded during calibration
    • roi_data.roi_mask, roi_data.roi_mask_bin, location of each neuron
    • roi_data.roi_bin_cell, mask of each neuron
    • roi_data.x, roi_data.y, location in x and y for each neuron
    • roi_data.im_roi, roi_data.im_roi_rg, images for plotting
• strcMask.mat: Structure mask data of the ensemble neurons used in the analysis. Relevant variables
  • strcMask.roi_ind, the index of each of the neurons.
  • strcMask.num_roi, number of Rois that do the ensemble
  • strcMask.maxx, location of the higher end in the Xaxis for each ensemble neuron. Needed to apply the neuronMask
  • strcMask.minx,  location of the lower end in the Xaxis for each ensemble neuron. Needed to apply the neuronMask
  • strcMask.maxy, location of the higher end in the Yaxis for each ensemble neuron. Needed to apply the neuronMask
  • strcMask.miny, location of the lower end in the Yaxis for each ensemble neuron. Needed to apply the neuronMask
  • strcMask.neuronMask, Mask for each ensemble neuron
  • strcMask.xctr, location of the center Xaxis for each ensemble neuron
  • strcMask.yctr, location of the center Yaxis for each ensemble neuron
  • strcMask.width, Width of each ensemble neuron
  • strcMask.height, Height of each ensemble neuron
• target_calibration.mat: Metadata used by the algorithm to calibrate the BMI.

Additional Information:

• Python: A high-level programming language widely used in scientific computing, data analysis, and machine learning. It is known for its readability and extensive libraries, which facilitate a broad range of applications in research and development.
• Pandas: A Python library used for data manipulation and analysis. It provides data structures like DataFrames.
• MATLAB: A programming platform designed for engineers and scientists. It allows for matrix manipulations, plotting of functions and data, implementation of algorithms, and creation of user interfaces. MATLAB is commonly used for data analysis and visualization in research.
• DataFrame: A two-dimensional, size-mutable, and potentially heterogeneous tabular data structure with labeled axes (rows and columns). It’s similar to a table in a database or an Excel spreadsheet.