bycycle.burst.recompute_edges¶

bycycle.burst.recompute_edges(df_features, threshold_kwargs, burst_method='cycles', burst_kwargs=None)[source]¶

Recompute the is_burst column for cycles on the edges of bursts.

Parameters:

df_featurespandas.DataFrame

A dataframe containing shape and burst features for each cycle.

threshold_kwargsdict

Feature thresholds for cycles to be considered bursts, matching keyword arguments for:

detect_bursts_cycles() for consistency burst detection (i.e. when burst_method == ‘cycles’)

Returns:

df_features_edgespandas.DataFrame: An cycle feature dataframe with an updated is_burst column for edge cycles.

Notes

df_features must be computed using consistency burst detection.

Examples

Lower the amplitude consistency threshold to zero for cycles on the edges of bursts:

>>> from neurodsp.sim import sim_combined
>>> from bycycle.features import compute_features
>>> sig = sim_combined(n_seconds=4, fs=1000, components={'sim_bursty_oscillation': {'freq': 10},
...                                                      'sim_powerlaw': {'exp': 2}})
>>> threshold_kwargs = {'amp_fraction_threshold': 0., 'amp_consistency_threshold': .5,
...                     'period_consistency_threshold': .5, 'monotonicity_threshold': .4,
...                     'min_n_cycles': 3}
>>> df_features = compute_features(sig, fs=1000, f_range=(8, 12),
...                                threshold_kwargs=threshold_kwargs)
>>> threshold_kwargs['amp_consistency_threshold'] = 0
>>> df_features_edges = recompute_edges(df_features, threshold_kwargs)