A Comment on Data Shuffling
David J C MacKay and Christopher deCharms and Virginia R. de Sa
In the analysis of neuronal spike trains
in response to experimental stimuli, the question
of whether there is `cooperativity' in the
coding of information in the spike trains
has been widely addressed by the practice of
data shuffling: a measure of
mutual information between stimulus
and the spike trains is computed using the
real data, then the data is shuffled so
as to destroy cooperative coding, if it is there,
and the same mutual information measure is computed
for the shuffled data.
In this note we apply this procedure to
a simple model in which it is questionable
whether the encoding should be termed `cooperative',
and show that the test nevertheless gives a positive
result.
postscript (Cambridge UK).
postscript (Canada mirror).
David MacKay's:
home page,
publications.
bibtex file.
Canadian mirrors:
home page,
publications.
bibtex file.