# gridcells.analysis.info - Information-theoretical analysis¶

The info module contains routines related to information-theoretic analysis of data related to grid cells.

## Functions¶

 information_rate(rate_map, px) Compute information rate of a cell given variable x. information_specificity(rate_map, px) Compute the ‘specificity’ of the cell firing rate to a variable X.
gridcells.analysis.info.information_rate(rate_map, px)[source]

Compute information rate of a cell given variable x.

A simple algorithm devised by [R3]. This computes the spatial information rate of cell spikes given variable x (e.g. position, head direction) in bits/second.

Parameters: rate_map : numpy.ndarray A firing rate map, any number of dimensions. If units are in Hz, then the information rate will be in bits/s. px : numpy.ndarray Probability density function for variable x. px.shape must be equal rate_maps.shape I : float Information rate.

Notes

If you need information in bits/spike, you need to divide the information rate by the average firing rate of the cell.

The firing rate map, in positions that are valid within the arena, may contains NaN numbers. In that case, the firing rate in these positions in rate_map will be set to 0.

References

 [R3] (1, 2) Skaggs, W.E. et al., 1993. An Information-Theoretic Approach to Deciphering the Hippocampal Code. In Advances in Neural Information Processing Systems 5. pp. 1030-1037.
gridcells.analysis.info.information_specificity(rate_map, px)[source]

Compute the ‘specificity’ of the cell firing rate to a variable X.

Compute information_rate() for this cell and divide by the average firing rate of the cell. See [R4] for more information.

Parameters: rate_map : numpy.ndarray A firing rate map, any number of dimensions. px : numpy.ndarray Probability density function for variable x. px.shape must be equal rate_maps.shape I : float Information in bits/spike.

References

 [R4] (1, 2) Skaggs, W.E. et al., 1993. An Information-Theoretic Approach to Deciphering the Hippocampal Code. In Advances in Neural Information Processing Systems 5. pp. 1030-1037.