Add 5 new SSVEP classifiers and fix label handling

[bruAristimunha]

Summary

• Add 5 new SSVEP classifiers: SSVEP_itCCA, SSVEP_eCCA, SSVEP_TRCA_R, SSVEP_SSCOR, SSVEP_TDCA
• Fix critical label mismatch in all SSVEP classifiers (CCA, TRCA, MsetCCA + new ones)
• Fix filterbank numerical stability (SOS form, capped filter order)
• Fix np.reshape deprecation in ssvep_nakanishi.py and ssvep_wang.py

Label handling fix

All SSVEP classifiers (except MsetCCA) used X.event_id.keys() to get class labels, which returned generic integer strings ('1', '2', ...) instead of actual frequency labels from y ('9.25', '9.75', ...). This caused:

1. Wrong sinusoidal reference frequencies (e.g., 1 Hz instead of 9.25 Hz)
2. Wrong filterbank peak frequencies
3. Label encoding mismatch between classifier predictions and evaluation y labels

Now all classifiers use np.unique(y) and boolean masking for epoch selection. The benchmark also passes mne_labels=True for epoch-based pipelines.

Benchmark results (Nakanishi2015, WithinSession, 12-class)

Pipeline	Before	After
CCA	10.0%	92.5%
itCCA	~8.3%	98.0%
eCCA	~8.3%	99.4%
MsetCCA	43.2%	43.2%
TRCA	~8.3%	86.3%
TRCA-R	~8.3%	99.3%
SSCOR	~8.3%	99.2%
TDCA	~8.3%	99.0%

Test plan

• All 32 classification tests pass (20 passed + 12 xpassed)
• WithinSession benchmark on Nakanishi2015 verified
• CI checks

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.