EEP - avrprocess

avrprocess - Basic ERP Operations

Synopsis


-----------------------------------------------------------------------------
EEP 3.1  Max-Planck-Institute of Cognitive Neuroscience 1996-99
avrprocess 3.6   (OSF1 V4.0 alpha)            Wed Sep 15 13:30:27 1999
-----------------------------------------------------------------------------
  avrprocess -o <avr out> [-l <label>] [-c <color>]
             <process options> [<avr in> ... ]

common options:
  -o <avr out> output filename/output directory
  -l <label>   conditon label for avr out
  -c <color>   condition color for avr out
  -t <ms>      add <ms> to trigger offset for avr out

process options:
  --set        set header information
  --diff [-v]  calc average difference
               -v   conserve trial numbers and variances
  --cum        accumulate single trials
  --avr  [-w]  calc grand average
               -w   S/N ratio weighting

Description

avrprocess is a general command line interface to several processing options. It's main purpose is to use the shell capabilities to build a list of input files. This way is faster and more flexible than having a special configuration with a proprietary list generation for each job.

Each run of avrprocess generates exact one average output file. You have to specify a name for it and you can specify a new label and a new color(per default, this information is copied from the first input file).

The output average always contains the largest common channel subset and the largest common time range of all input averages.

Processes

--diff [-v]

calc average difference(first file - second file)
exact two input files required
normally, the output avr will contain no variance info; the number of trials is set to 1, the number of rejected trials is set to 0
with the -v option, the variances are copied from the first operand and the trial numbers are unchanged - use with care, this is normally an invalid operation

--set

set header information
exact one input file required
no data changes; output contains new label and/or color only

--cum

accumulate single trials
output contains the average of the single trials of the input averages(mainly used to merge blockwise calculated averages into one full-session average)

--avr [-w]

calc grand average
output contains the mean and variance of the input averages; optionally, you can weight the input files with sqrt(N_Trials)

Examples (bash required)

avrprocess -o vp01r.avr -l rare_all -c green --cum vp01r{1,2,3}.avr

accumulates the single trials of the files

vp01r1.avr
vp01r2.avr
vp01r3.avr

- assume they contain blockwise ERP's of a condition "rare" - into the output average

vp01r.avr

which then contains the ERP of the rare condition for the whole session