v0.3.0 — Full Matlab Parity

What's New in v0.3.0

This release achieves full method-level parity with the Matlab nSTAT toolbox across all 16 audited classes (484 methods).

New Methods

• SignalObj: MTMspectrum(), spectrogram(), periodogram(), findPeaks(), findMaxima(), findMinima(), findGlobalPeak(), shift(), shiftMe(), alignTime(), power(), sqrt(), xcov()
• DecodingAlgorithms: Full SSGLM EM suite (PPSS_EStep, PPSS_MStep, PPSS_EM, PPSS_EMFB, estimateInfoMat, prepareEMResults), UKF methods (ukf, ukf_ut, ukf_sigmas)
• SpikeTrainCollection: ssglm(), ssglmFB() entry points
• FitResSummary: plotAllCoeffs(), plot3dCoeffSummary(), plot2dCoeffSummary(), plotKSSummary()

Paper Examples

All 5 paper examples rewritten as self-contained, documented scripts matching the Matlab workflow:

1. mEPSC Poisson analysis (constant vs piecewise-rate)
2. Whisker stimulus thalamus (lag search, nested model comparison)
3. PSTH and SSGLM (CIF simulation, GLM-PSTH, state-space learning)
4. Place cells (Gaussian + Zernike receptive fields, population statistics)
5. Decoding (PPAF, goal-directed reach, hybrid filter)

Fidelity Summary

• Classes: 11 exact, 7 high fidelity, 1 N/A (19 total)
• Notebooks: 8 exact, 5 high fidelity (13 total)
• Tests: 198 passed, 2 skipped

Install

pip install nstat-toolbox==0.3.0

Full changelog: v0.2.0...v0.3.0