1.4.0

First modern release after the rename from pyEEG to natmeeg, consolidating
years of development into a single tagged milestone.

Added

• Temporal Response Functions: TRFEstimator with SVD-based ridge
  regression, multi-alpha scoring, list-based covariance accumulation,
  load/save, xfit cross-validation, interactive and significance plotting,
  topomaps, and __add__/__truediv__/__getitem__ operators.
• Canonical Correlation Analysis: CCA_Estimator (lagged/regularized)
  plus cca_nt and cca_svd backends; multiway CCA (mCCA) for
  hyperalignment preprocessing.
• Connectivity module: Granger causality, phase transfer entropy (PTE),
  phase linearity measurement (PLM), and cross-spectral density.
• Simulation module: AR/VAR generation, Jansen-Rit (and extended) neural
  mass model, CTRNN, and TRF simulation kernels.
• Preprocessing: Whitener (PCA/ZCA), WaveletTransform,
  MultichanWienerFilter (MWF), filterbanks, and covariance estimators.
• IO: EEGLAB/FieldTrip → MNE conversion, AlignedSpeech and
  WordLevelFeatures classes, word-onset/envelope/surprisal/syntactic
  feature loaders.
• Visualization (pyeeg.vizu): topomaps, filterbank plots, TRF
  significance overlays, pairwise boxplots, interactive plots.
• Gammatone filterbank and cochleagram C extensions.
• Sigstore-signed PyPI publishing via GitHub Actions.

Changed

• MNE made an optional dependency; psutil and tqdm made required.