psth

Make plots of mean firing rate across trials as a function of time relative to a marker, e.g., stimulus onset.

Options

The output type (on the third line of the .nda file) can be

• simple - plain plot of PSTH values.
• simple_matlab - single column of float values
• histogram - writes a plot file with three points that form a histogram box for every data bin.
• power - compute the power spectrum of the PSTH using data-windowing. Uses Numerical Recipes in C "spctrm" routine, where 'ovrlap' is set to 1. The routine will use as many overlapping (by half the window length) data windows as fit into the length (in bins) of the PSTH, and will discard the rest of the PSTH.

Parameters

• start - start time for binning (sampling units).
• period - duration of plot (sampling units).
• binsize - width of histogram bins, default is 20 (sampling units) for outtype histogram, 1 for simple. The total number of bins is determined by period/binsize.
• sigma - SD for Gaussian smoothing, default is 0.0 (no smoothing), in binsize units.
• toffset - added to PSTH time axis.
• yscale - multiply PSTH value by this before plotting.

  Variable start time

• var_start_param - parameter name for variable start time value
• var_start_mult - multiply the variable start value by this number
• var_start_add - add this to variabe start value after multiplying

  POWER

• pwr_win_size - (128) window size for power spectrum, must be a power of two. Default is 128. This will determine the number of bins in the spectrum.
• pwr_zero_mean - (0) set to 1 to subtract the mean from the PSTH

Examples are given below.

Example

#  
#  Post-stimulus time histogram.
#
psth
all
simple

start 0
period 2000
binsize 1
sigma 2.0

Example

#  
#  Post-stimulus time histogram.
#
psth
group 0
histogram

start 0
period 2000
binsize 100

Example

#
#  Power spectrum of PSTH
#
psth
all
power

binsize 5         # PSTH binsize (affects the *range* of the frequency axis)
                  #    e.g, 5ms = 200Hz, thus a 0-100 Hz frequency range
                  #    will apply to the resulting power spectrum

pwr_win_size 256  # Window size (this affects *frequency* binsize
                  #  but *not* the frequency range
                  #  E.g., there will be 256 bins from 0 to cutoff, thus, your
                  #  freq. resolution (or power binsize) is 100Hz/256 = 0.39Hz

pwr_zero_mean 1   # 1-Subtract mean from PSTH;  0-Do not subtract mean.

Example

psth
group 1 stn
simple

start -500
period 16000

var_start_param stn
var_start_mult  1000.0   # Convert seconds to msec
var_start_add    100.0   # Account for st0 = 100ms (added after mult)

binsize 1
sigma   4.0