Parameters struct (P)

Parameter file format

Recording file format

Name	Default Value	Description
vcFile	''	Raw recording file path (leave empty if multiple files are sorted together)
csFile_merge	{}	List of recording files to merge (leave empty if single file is used)
sort_file_merge	1	Sort files by the 0: alphabetical order, 1: metafile, 2: file created time, 3: file modified time, 4: numerical order in the file names
vcFile_prm	''	Current parameter file path (.prm) if set empty
template_file	''	Template parameter file path (.prm). Parameters in this file override the default parameter file (default.prm), but get overridden is by the parameter file.
probe_file	''	Probe file to use (.prb). Default search location: ./prb/
nChans	120	# channels stored
vcDataType	'int16'	Raw data format. Choices: {'int16', 'uint16', 'int32', 'uint32', 'single', 'double'}
sRateHz	25000	Sampling rate (Hz)
uV_per_bit	0.305176	Bit scaling factor (uV/bit)
fTranspose_bin	1	Set to 1 if the samples from the same time step is written to the file contiguously.
fInverse_file	0	Set to 1 to flip the polarity of the signal. Set to 0 if spikes have negative going polarity.
header_offset	0	File header offset (in bytes). Set to 0 if there is header info (e.g. WHISPER format)

Execution parameters

Name	Default Value	Description
version	'v3.x.x'	JRCLUST version. Updated on "month" "day", "year"
fVerbose	0	Verbose flag, set to 0 to suppress displaying extra information
fParfor	1	Use Multiple CPU cores (if parallel processing toolbox is installed)
fGpu	1	Use GPU if parallel processing toolbox is installed
iGpu	1	Use specific GPU ID (run 'gpuDevice()' to retrieve the list of GPUs)
MAX_BYTES_LOAD	[]	Default memory loading block size (bytes)
MAX_LOAD_SEC	[]	Maximum loading duration (seconds). This overrides "MAX_BYTES_LOAD".
MAX_LOG	10	Maximum number of histories to track for 'manual' GUI
nSamples_gpu	250000	Number of samples to process in GPU at a time. Reduce this number if GPU crashes for high-channel arrays
vcFile_thresh	''	Name of .mat file taht stores the spike detection threshold in 'vnThresh_site' variable. Created by 'preview' GUI
sec_per_load_preview	1	Recording duration (sec) per continuous segment to preview
nLoads_max_preview	30	Number of time segments to load for preview
offset_sec_preview	0	Time offset (sec) to apply to the preview
fCacheRam	1	Cache spike waveform information in RAM. Uses more RAM but faster
fRamCache	1	Cache both raw and filtered spike waveforms in RAM (greater memory demand)

Display parameters (Affects display only, (*):Can be modified in the manual GUI)

Name	Default Value	Description
tlim	[0, .2]	Time range to display
nTime_traces	1	Number of time segments to display. Set to 1 to show one continuous time segment
nSkip_show	1	Show 1 in n samples when plotting traces
fSpike_show	1	show spikes in the trace view (*)
maxAmp	250	Amplitude range in uV (initial setting, changeable in UI) (*)
nShow	200	Maximum number of traces to show [D?# spikes to show]
corrLim	[.9 1]	Correlation color plot range
nSpk_show	30	show spike waveforms for manual clustering
nShow_proj	500	Max. number of spikes to show in the projection view
vcFet_show	'vpp'	Feature to show in time plot {'vpp', 'cov', 'pca'} (*)
mrColor_proj	[.75 .75 .75; 0 0 0; 1 0 0]	background, selected, comparison color
vcFilter_show	''	Filter to apply in the "traces" view. Choices: {'', 'bandpass', 'sgdiff'}. '' uses vcFilter setting.

Pre-processing (run "jrc spikesort" after change)

Name	Default Value	Description
tlim_load	[]	Time range to load in sec (uses all time range if set to empty)
vcFilter	'ndiff'	Filter types: {'ndiff', 'sgdiff', 'bandpass', 'fir1', 'user', 'none'}
nDiff_filt	2	Differentiation filter for vcFilter='sgdiff', ignored otherwise. Set to [] to disable. 2n+1 samples used for centered differentiation
freqLim	[300, 3000]	Frequency cut-off limit for vcFilter='bandpass', ignored otherwise.
filtOrder	3	Bandpass filter order
fEllip	1	Use Elliptic filter. set to 0 for Butterworth filter. Only valid for vcFilter='bandpass'.
vnFilter_user	[]	User-specified filter kernel. Used if vcFilter='user' is set.
fft_thresh	0	Automatically remove frequency outliers (unit:MAD, 10 recommended, 0 to disable). Verify by
freqLimNotch	{}	Notch filter frequency range to exclude manually. {[start1, end1], [start2,
vcCommonRef	'mean'	common reference mode {'none', 'mean', 'median', 'whiten'}
viSiteZero	[]	Site# to ignore manually
nPad_filt	100	Number of samples to overlap between multiple loading (filter edge safe).
gain_boost	1	multiply the raw recording by this gain to boost uV per bit
thresh_corr_bad_site	[]	Automatically reject bad sites based on the max. correlation with neighboring sites using raw waveforms. Set to [] to disable

Spike detection and grouping (run "jrc spikesort" after change)

Name	Default Value	Description
qqFactor	5	Spike detection thresold (Threshold = qqFactor*S, S = med(x-med(x))/.6745). (RQ Quiroga'04)
qqSample	4	Median subsample factor (1 every n)
spkThresh_uV	[]	Manual threshold (applies to all channels)
spkThresh_max_uV	[]	maximum absolute amp. allowed
spkLim_ms	[-.25, .75]	Time range to capture the filtered spike waveforms
spkLim_raw_ms	[]	Raw spike waveform range, setting to [] will use spkLim_ms
maxSite	[]	Auto-calculated by `maxDist_site_spk_um` if empty. `nSites_spk` = 1+2*`maxSite`-`nSite_ref`)
spkRefrac_ms	.25	Detection refractory period in milliseconds
vcSpatialFilter	'none'	Spatial filter for detection {'mean', 'imec2', 'subtract', 'csd'}
fSaveEvt	1	Flag to save event file
fDetectBipolar	0	bipolar amplitude detection (0 means negative)
fSaveRawSpk	0	Save spike detected per site (saves minimum peak)
blank_thresh	[]	reject spikes exceeding the channel mean after filtering (MAD unit), ignored if [] or 0
blank_period_ms	5	(miliseconds) Duration of blanking when the common mean exceeds a threhold (blank_thresh)
nneigh_min_detect	0	Min. number of neighbors near the spike below threshold. choose between [0,1,2]
maxDist_site_um	50	Radius of spike event merging. Keep spikes that has most negative peaks within x um radius.
maxDist_site_spk_um	70	Max. radius for extracting waveforms. Used if maxSite=[] and nSites_ref=[]

Feature extraction (run "jrc spikesort" after change)

Name	Default Value	Description
vcFet	'gpca'	Feature list: {'gpca', 'cov', 'vpp', 'vmin', 'vminmax', 'energy', 'pca'}
nPcPerChan	1	number of principal components per chan.
nSites_ref	[]	set to 0 for global referencing (move to feature section)
vcSpkRef	'nmean'	Subtract mean of the spiking event sites {'nmean', 'none'}
fInterp_fet	1	Interpolate the waveforms for feature projection to find the optimal delay (2x interp). Set to 0 to disable

Clustering (run "jrc sort" after change)

Name	Default Value	Description
nTime_clu	4	Number of time periods to perform separate clustering (later to be merged after clustering)
dc_percent	2	used for distribution cutting
fRepeat_clu	0	Flag for whether to repeat clustering for the bottom half of the cluster amplitudes
nThreads	128	number of GPU threads for clustering
nC_max	45	(read-only) Set by the CUDA code. Maximum number of dimensions per event
fSpatialMask_clu	0	Apply spatial mask calculated from the distances between sites to the peak site (half-scale: P.maxDist_site_um). circa v3.1.10
min_sites_mask	5	Minimum number of sites to have to apply spatial mask. Used if fSpatialMask_clu==1. circa v3.1.10

Post-cluster auto-merging parameters (run "jrc auto" after change)

Name	Default Value	Description
delta1_cut	.6	RL cluster: delta cut-off. Increasing this will produce less clusters.
rho_cut	-2.5	RL cluster: rho cut-off. Decreasing (more negative) this will produce more clusters.
maxWavCor	.98	merge if waveform correlation is greater than maxWavCor (set to 1 to disable).
min_count	30	Minimum cluster size (min_count will be set to 2*#features if lower).
nRepeat_merge	10	# of repeats for average unit raw waveform based merging
vcDetrend_postclu	'global'	'none', 'hidehiko', 'global', 'local', 'logz'
thresh_mad_clu	7.5	This sets the MAD score threshold to remove outlier spikes for each cluster.
maxCluPerSite	20	Maximum clusters per site allowed if local detrending is used.
fDrift_merge	1	Compute multiple waveforms at three drift locations based on the spike position (CoM).
nInterp_merge	1	Interpolation factor for the mean unit waveforms, set to 1 to disable
maxDist_site_merge_um	35	Max. distance separation of the peak sites to consider merging
spkLim_factor_merge	1	Waveform range for computing the correlation. spkLim_factor_merge <= spkLim_raw_factor_merge. circa v3.1.8

Trial

Name	Default Value	Description
vcFile_trial	''	.mat or .csv file containing timestamp in seconds unit. use any variable name.
thresh_trial	[]	Threshold to detect trial event timing from the TTL pulse. If empty, half between min and max is used.
tRefrac_trial	.001	Trial timestamp refractory period (sec), disable duplicate events
tlim_psth	[-1 5]	Time range to display PSTH (in seconds)
tbin_psth	.01	Time bin for the PSTH histogram (in seconds)
xtick_psth	.2	PSTH time tick mark spacing (in seconds)
fAverageTrial_psth	1	plot individual trials
nSmooth_ms_psth	50	PSTH smoothing time window (in milliseconds)
rateLim_psth	[]	range range in Hz

Firing rate estimation for unit

Name	Default Value	Description
sRateHz_rate	1000	Resampling rate for the firing rate
filter_sec_rate	2	Time period to determine the firing rate
filter_shape_rate	'triangle'	{'triangle', 'rectangle'} kernel shape for temporal averaging

Aux channel

Name	Default Value	Description
vcFile_aux	''	Aux channel file
sRateHz_aux	[]	sampling rate for aux file
iChan_aux	[]	Aux channel # to correlate with the unit firing rate
vcLabel_aux	''	Label for the aux channel
vrScale_aux	1	Scale factor for the aux input
nClu_show_aux	10	Max # clusters to show for aux vs. firing rate correlation