Model

Module for algorithm computing

Bayesian Position decoding

Position decoding using population neuronal activity

• Refer to API: neuralib.model.bayes_decoding

RasterMap

RasterMap is an unsupervised discovery algorithm for neural data, developed by Carsen Stringer and Marius Pachitariu. It is particularly useful for identifying spatial or temporal patterns in large-scale neural recordings.

• Refer to API: neuralib.model.rastermap

This module provides the following functions:

• run: Perform RasterMap embedding on neural or imaging data.

• plot: Visualize RasterMap outputs, optionally alongside behavioral measurements.

• read: Load previously saved RasterMap results.

Supported Input Data

• Cellular datasets (e.g., electrophysiology spikes or calcium imaging):

  Array[float, [N, T]]

  • N: Number of neurons

  • T: Number of timepoints or samples

• Widefield imaging datasets:

  Array[Any, [T, H, W]]

  • T: Number of timepoints

  • H: Image height (pixels)

  • W: Image width (pixels)

References

• GitHub repository

• Example notebook (cellular data)

• Example notebook (widefield data)