phaseconsistency

Phase consistency refers to the stability of the phase of a time-varying signal across repetitions, trials, or channels. It measures how consistently the phase angle at a given frequency and time point aligns across observations, independent of changes in amplitude. In many fields, particularly neuroscience, phase consistency is used as an indicator of synchronized neural activity in response to a stimulus or task.

Measuring phase consistency typically involves extracting the instantaneous phase of the signal at a chosen frequency and time using methods such as the Hilbert transform or wavelet analysis. For a set of trials, the phase at each time–frequency point is combined to produce a measure of inter-trial phase coherence or phase-locking value. A common formulation is the mean resultant length, ITPC or PLV, computed as the magnitude of the average unit phasors across trials: ITPC = (1/N) |∑ e^{i φ_k}|, yielding a value between 0 and 1, where 1 indicates perfect phase alignment and 0 indicates random phase across trials. Phase consistency can also be described via circular statistics such as circular variance.

Interpretation of results depends on experimental context: high phase consistency suggests that the observed event reliably drives timing of neural or system responses, while low consistency implies variability in phase or timing. Analyses must account for biases from small trial counts, nonstationarities, and potential confounds like volume conduction or noise. Permutation testing and appropriate corrections help assess statistical significance.

Related concepts include phase synchronization, coherence, and other phase-based metrics used across signal processing and neuroscience to quantify temporal alignment and coordination.