Date started: April 2014 Date ended: ACTIVE

This project analyzes the timing of alternating mouse button presses (right/left or a/b) made in response to a bi-stable visual stimulus.

main.m will (1) get data, (2) generate a button press time series (list of mouse states as a function of time with a fixed dt), (3) summarize and visualize the time series pre-clean-up, (4) clean-up overlaps (not yet gaps) in button presses, (5) summarize and visualize time series post-clean-up, and (6) analyze the time series data.

Analyze the responses of experimental subjects in a vision science experiment.

• add option to remove 1st epoch in deriveVars.m
• clean-up plots
• clean-up gaps
• remove dominance durations < 300 ms; see, 'Plaid-200415_MR_glassesWfilter_manipulatedLambdasB.txt_trial_2'
• analysis section for group data

• config.m - set parameters including working directories/paths to data (Written by EB)
• main.m - entry point for running analysis script (Written by EB)

Open config.m and (a) set the path to select data, (b) set the path to store data and results, and (c) set the value of the variable "analysis_type". "analysis_type" can be set equal to 1 to analyze a single file; equal to 2 to analyze all subject data saved in keyData.mat; set equal to 3 to analyze simulated dataset. (d) Also, set the value for "input_format_vers", e.g., the version of the raw data (1 = old format; 2 = new format). Then run main.m.

• importfile.m - Import numeric data and string data from a text file as a matrix

  • INPUT: filename
  • OUTPUT: Data, splitdata, numTrials, index
    • Data (7 columns + n columns for n trials)
      • column 1 = time of left/right button press/release
      • column 2 = NaN's
      • column 3 = NaN's
      • column 4 = similar to column 5; not used
      • column 5 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
      • column 6 = NaN's
      • column 7 = NaN's
      • column 8 = Trial 1 list of all configuration and trial parameter values
    • splitData (1 x n structure, with n = # trials)
      • trial 'n' (4 fields; data, name = filename + trial#, params, paramsNames
        • data (3 columns) -
          • column 1 = count from 1 : length(column 2)
          • column 2 = time of left/right button press/release
          • column 3 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
        • params (1 column) - list of all configuration and trial parameter values
        • paramsNames (1 column) - list of all config and trial parameter names

• simTC_wJitter.m - simulate pseudo-random button press event series

  • SUMMARY: Simulates pseudo-random event series of alternations between percept A and B. Percept durations are drawn from a log-normal distribution. Called by main.m.
  • INPUT: meanAmsec, stdAmsec, meanBmsec, stdBmsec, meanRT, RTj, N
    • The average ('mean') and std of percepts A and B are specified independently
    • meanRT = mean reaction time
    • RTj = jitter in reaction time (standard deviation)
    • N = number of simulated durations for each percept
  • OUTPUT:
    • simTC (3 columns)
      • column 1 = count from 1 : length(column 2)
      • column 2 = time of left/right button press/release
      • column 3 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
  • (Written by NR)

• NB: behavioral data is imported via main.m

• selectData.m -

  • SUMMARY: This function imports the .txt or .dat raw data files. Called by main.m.
  • INPUT: formatType(#), st_dir(string), out_dir(string), raw_mat(string)
  • OUTPUT: numFiles(#), keyData, index(scalar for format 1; vector for format 2)
    • keyData (3 columns)
      • column 1 = count from 1 : length(column 2)
      • column 2 = time of left/right button press/release
      • column 3 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
  • (Written by EB)

• sepTrials.m -

  • SUMMARY: Raw data in format type #2 has multiple trials in one file. This function separates the button press data into those multiple trials. Called by the importfile.m function.
  • INPUT: rawData(7+ column matrix), filename (string)
  • OUTPUT: trialData (1 x n structure, with n = # trials)
    • trial 'n' (4 fields; data, name = filename + trial#, params, paramsNames
      • data (3 columns) -
        • column 1 = count from 1 : length(column 2)
        • column 2 = time of left/right button press/release
        • column 3 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
      • params (1 column) - list of all configuration and trial parameter values
      • paramsNames (1 column) - list of all config and trial parameter names
  • (Written by EB)

• genTimeSeries.m -

  • SUMMARY: raw data is in the format of a column of mouse press/release events and needs to be converted to an evenly sampled time series format. Called by main.m.
  • INPUT: rawData, framerate (#), endTime(#)
    • rawData (3 columns)
      • column 1 = count from 1 : length(column 2)
      • column 2 = time of left/right button press/release
      • column 3 = label of event; Aon = 1; Aoff = 2; Bon = -1; Boff = -2
  • OUTPUT: timeSeries (3 columns), endTime(#) - column 1 = time at 120 Hz - column 2 = label of A events; Aon = 1, Aoff = 0 - column 3 = label of B events; Bon = 1, Boff = 0
  • (Written by EB in conjunction with NR)

• addLastRelease.m -

  • SUMMARY: sometimes a button is still pressed at the end of a trial; this function adds a dummy 'release' to the last frame of the trial; the function returns button press vectors of equal length. Called by genTimeSeries.m.
  • INPUT: vecAon, vecAoff, vecBon, vecBoff, ES, timeMaxTS(#)
    • vecAon/vecAoff/vecBon/vecBoff (1 column)
      • column 1 = indices of that button event in ES
    • ES (3 columns) == rawData (see above)
  • OUTPUT: vecAoff, vecBoff, ES
    • vecAoff/vecBoff (1 column)
      • column 1 = indices of that button event in ES, plus one additional index for either vecAoff or vecBoff (in effect, to add last release)
    • ES (3 columns) - new row added at end
      • column 1 = last cell set as len(col) + 1
      • column 2 = last cell set as timeMaxTS (last timestamp in trial)
      • column 3 = last cell set equal to either 2 (if add last release to vecAoff) or -2 (if add last release to vecBoff)
  • (Written by EB adapted from NR's 'analyzeTC.m')

• findPressInd.m -

  • SUMMARY: the time series vector needs to be labeled with the corresponding button press events. Events are labeled for each frame with a 1 if the button was pressed or 0 if a button was not pressed (press A = column 2; press B = column 3). This function finds the times when a button was pressed (pressOn) and released (pressOff) and returns the index of the nearest frame at 120Hz. Called by genTimeSeries.m.
  • INPUT: vecPressOn, vecPressOff, i (#), ES, FR(#)
    • vecPressOn (1 column)
      • column 1 = indices of that button event (see vecAon, vecBon)
    • vecPressOff (1 column)
      • column 1 = indices of that button event (see vecAoff, vecBoff)
    • ES (3 columns) == rawData (see above)
  • OUTPUT: indStart (#), indEnd (#)
  • (Written by EB in conjunction with NR)

• sortData.m - POSSIBLY DELETE - simple sorting function (Written by EB)

• cleanUpTS.m

  • SUMMARY: Find indices when col2 and col3 of TS are both pressed (==1). Loop backward (check last overlap 1st). If 1 frame after that overlap for pressA is equal to zero, change that overlap frame for pressA to zero. Else, pressB for that overlap frame is changed to zero. Called by main.m.
  • INPUT: TS (== timeSeries, see above)
  • OUTPUT: TS (3 columns)
    • column 1 = no change
    • column 2 = change 1's to 0's for press A if meets conditions in summary
    • column 3 = change 1's to 0's for press B if meets conditions in summary
  • (Written by EB)
  • NB: TO-Do; this function clean-ups overlaps but not gaps

• summarizeData.m -

  • SUMMARY: this function takes the timeSeries data (button press as a function of time at 120 Hz) and returns (1) the length of each gap/overlap between press A and B, (2) the duration of each press, and (3) the # of press switches. Called by main.m.
  • INPUT: TS, filename (string), plot_yn (#), startTime(#)
    • TS (3 columns) (== timeSeries, see above)
      • column 1 = time at 120 Hz
      • column 2 = label of A events; Aon = 1, Aoff = 0
      • column 3 = label of B events; Bon = 1, Boff = 0
  • OUTPUT: gapOverlap, meanGapOverlap (#), stdGapOverlap (#), durA, durB, numSwitch (#)
    • gapOverlap (2 columns)
      • column 1 = duration of each gap ('pos' #) or overlap ('neg' #)
      • column 2 = label of each event; A to B = -2; B to A = 2; repeat button (same press) = 0
    • durA (1 column)
      • column 1 = duration of each A press
    • durB (1 column)
      • column 1 = duration of each B press
  • (Written by EB, adapted from 'analyzeTC.m' written by NR)

• deriveVars.m -

  • SUMMARY: this function inputs the timeSeries data and returns (1) the % time of percept A and B, (2) the mean dominance durations of press A and B, (3) the reaction time to the first button press (s), and (4) the rate of button press alternations (s). Called by main.m.
  • INPUT: TS (see TS above), numSwitches (#), domDurA (see durA above), domDurB (durB above), FR(#)
  • OUTPUT: relA (#), relB (#), meanDurA (#), meanDurB (#), RT (#), altRate (#)
  • (Written by EB)

• analyzeTC.m - returns timestamps of button presses (Written by NR)

• plotTC.m - plot time course of button presses
  • Called by summarizeData.m.
  • (Written by NR)
• histogramPrep.m - return hist vectors for mouse press durations
  • Called by summarizeData.m.
  • (Written by EB, adapted from NR)

• crossCorrelation.m
• autocorrelation.m