nerve_response_analysis.analysisRun.measureCapMetrics

Description

Measures baseline, noise, positive and negative response peaks, peak-to-peak amplitude, dominant-peak latency, and signal-to-noise ratio for each stimulus time. All time options and outputs are in seconds. Signal-derived values retain the units of signal; SNR is reported in decibels.

Inputs

timeSec

Numeric sample-time vector in seconds.

signal

Numeric response vector with the same length as timeSec.

eventTimesSec

Numeric vector of stimulus times in seconds. One output row is produced for each value, in input order.

opts

Optional scalar structure containing the measurement windows below.

Options

segments

Optional scalar protocol structure. segments.capSearch may contain blankingAfterPulseSec and searchEndAfterPulseSec; nested values override direct options. Default: struct().

baselineWindowSec

Duration before each event used to estimate baseline. The inclusive baseline interval is [event-baselineWindowSec, event-blankingAfterPulseSec]. Default: 0.050 seconds.

blankingAfterPulseSec

Time excluded immediately before and after the stimulus. It may also be supplied as opts.segments.capSearch. blankingAfterPulseSec. The nested value takes precedence. Default: 0.002 seconds.

searchEndAfterPulseSec

End of the inclusive response-search interval, measured after the event. It may also be supplied in opts.segments.capSearch and takes the same precedence. Default: 0.008 seconds.

Calculations

baselineMean is the finite-value mean in the baseline interval. noiseRms is the root mean square deviation from that mean. If the interval contains no samples, the recording-wide finite median becomes the baseline and noiseRms remains NaN. The response search is [event+blankingAfterPulseSec, event+searchEndAfterPulseSec]. Peaks are measured after baseline subtraction. peakTimeSec belongs to whichever of the positive or negative peaks has larger absolute amplitude, and latencySec = peakTimeSec-stimTimeSec. When noiseRms is positive and finite, snrDb = 20*log10(abs(peakToPeak)/noiseRms).

Outputs

metrics

Table with one row per event.

Metric Columns

eventIndex - One-based event order. stimTimeSec - Input stimulus time. baselineMean, noiseRms - Pre-stimulus baseline and noise estimate. peakPositive, peakNegative, peakToPeak - Baseline-subtracted extrema and their difference. peakTimeSec, latencySec - Dominant absolute peak time and event-relative latency. snrDb - Peak-to-peak SNR in decibels, or NaN when noise is unavailable or zero. status - "ok" when the search interval contains samples; "noSamples" otherwise. Rows without samples retain NaN response metrics.

Errors

nerve_response_analysis:MetricSizeMismatch

timeSec and signal lengths differ.

Example

timeSec = (0:0.0001:0.04).';
signal = zeros(size(timeSec));
signal(abs(timeSec-0.014) < 0.0002) = 2;
signal(abs(timeSec-0.016) < 0.0002) = -1;
opts = struct("baselineWindowSec", 0.004, ...
"blankingAfterPulseSec", 0.002, ...
"searchEndAfterPulseSec", 0.008);
metrics = nerve_response_analysis.analysisRun.measureCapMetrics( ...
timeSec, signal, 0.010, opts);
assert(metrics.status == "ok" && metrics.peakToPeak > 2.5)