A couple of Matlab classes to make handling spike sorting data substantially easier.
SingleUnit.m: class to store individual single units, with useful methods to inspect waveforms, autocorrelations etc.
MultipleUnits.m: class to store populations of single units, with methods to assess population firing activities.
A more thorough example of usage, starting from the raw output of UltraMegaSort is in example_usage.m.
Quickstart:
data = MultipleUnits('patient', 'JoeBloggs', 'seizure', 1);
unit = SingleUnit('times',spiketimes,'waveforms',waves,'channel',1);
data.addUnit(unit);
% add all desired units (usually in a loop over output files)
data.order_by_rate();
data.raster();Selected properties:
| Property | Description |
|---|---|
| UID | Unique ID for this neuron |
| waveforms | [m by n] matrix of m spikes across n datapoints |
| times | Column vector of spike times in your preferred units (e.g. s, ms, datapoints) |
| wideband | Mean waveform from the broadband signal (used for cell-type subclassification) |
| type | Type of neuron (e.g. FS-IN, RS, PC) |
| extra | Structure for storage of extra data not core to the object |
Selected methods:
| Method | Description |
|---|---|
| inspect_unit | Plots an overview of this unit for visual inspection |
| autocorr | Calculate the autocorrelation for this unit |
| xcorr | Calculate the cross-correlation between this and another SingleUnit object |
| gaussian_fr | Estimate the instantaneous firing rate of this unit with a Gaussian kernel |
| hist_fr | Calculate the binned firing rate of this unit |
| mean_ac_lag | Calculate the mean lag of this unit's autocorrelation (for subclassification) |
| ISI | Calculate the ISI for this unit (autocorrelation is better than raw ISI...) |
| plot_* | Plot the output of various other calculations |
| retrieve_* | Return the values of the various other calculations |
Properties:
| Property | Description |
|---|---|
| patient | Patient identifier (string) |
| seizure | Seizure number (int) |
| epoch | Start and finish times of epoch [double double] |
| units | Array of SingleUnit objects |
| snr | Stores the signal-to-noise ratio of this recording once calculated |
| info | Any extra details and info for this recording |
Selected methods (see example_usage.m for input/output explanations):
| Method | Description |
|---|---|
| add_unit | Add a SingleUnit object to this MultipleUnits object |
| all_spike_times | Return all spike times across all units within specified epoch (defaults to all) |
| beefy_raster | Make a comprehensive raster plot of these units, allowing color-coding. Not recommended: User .raster() for speed. |
| channel_units | Return array of all SingleUnit objects from specified channel. |
| gaussian_fr | Calculate the Gaussian estimate of the population firing rate across all units within this MultipleUnits |
| order_by_channel | Order the SingleUnits by channel number |
| order_by_rate | Order the SingleUnits by overall firing rate |
| plot_channel_units | Plot all units from specified channel to assess separation accuracy (see screenshot below) |
| raster | Make a basic raster plot of these units. No color-coding, not very pretty - use .beefy_raster() for comprehensive plot |
| unit_snr | Calculate the SNR across all units in the object |
| top_channels | Return the specified number of channels with the most units recorded |
Example output of the MultipleUnits plot_channel_units(channel) method:
