py-neuromodulation 0.0.4__py3-none-any.whl → 0.0.5__py3-none-any.whl

py_neuromodulation/nm_oscillatory.py

@@ -1,448 +1,423 @@
-from typing import Iterable
-
-import numpy as np
-from scipy import fft, signal
-
-from py_neuromodulation import nm_filter, nm_features_abc, nm_kalmanfilter
-
-
-class OscillatoryFeature(nm_features_abc.Feature):
-    def __init__(
-        self, settings: dict, ch_names: Iterable[str], sfreq: float
-    ) -> None:
-        self.s = settings
-        self.sfreq = sfreq
-        self.ch_names = ch_names
-        self.KF_dict = {}
-
-        self.f_ranges_dict = settings["frequency_ranges_hz"]
-        self.fband_names = list(settings["frequency_ranges_hz"].keys())
-        self.f_ranges = list(settings["frequency_ranges_hz"].values())
-
-    @staticmethod
-    def test_settings_osc(
-        s: dict,
-        ch_names: Iterable[str],
-        sfreq: int | float,
-        osc_feature_name: str,
-    ):
-        assert (
-            fb[0] < sfreq / 2 and fb[1] < sfreq / 2
-            for fb in s["frequency_ranges_hz"].values()
-        ), (
-            "the frequency band ranges need to be smaller than the nyquist frequency"
-            f"got sfreq = {sfreq} and fband ranges {s['frequency_ranges_hz']}"
-        )
-
-        if osc_feature_name != "bandpass_filter_settings":
-            assert isinstance(
-                s[osc_feature_name]["windowlength_ms"], int
-            ), f"windowlength_ms needs to be type int, got {s[osc_feature_name]['windowlength_ms']}"
-
-            assert (
-                s[osc_feature_name]["windowlength_ms"]
-                <= s["segment_length_features_ms"]
-            ), (
-                f"oscillatory feature windowlength_ms = ({s[osc_feature_name]['windowlength_ms']})"
-                f"needs to be smaller than"
-                f"s['segment_length_features_ms'] = {s['segment_length_features_ms']}",
-            )
-        else:
-            for seg_length in s[osc_feature_name][
-                "segment_lengths_ms"
-            ].values():
-                assert isinstance(
-                    seg_length, int
-                ), f"segment length has to be type int, got {seg_length}"
-        assert isinstance(
-            s[osc_feature_name]["log_transform"], bool
-        ), f"log_transform needs to be type bool, got {s[osc_feature_name]['log_transform']}"
-
-        assert isinstance(s["frequency_ranges_hz"], dict)
-
-        assert (
-            isinstance(value, list)
-            for value in s["frequency_ranges_hz"].values()
-        )
-        assert (len(value) == 2 for value in s["frequency_ranges_hz"].values())
-
-        assert (
-            isinstance(value[0], list)
-            for value in s["frequency_ranges_hz"].values()
-        )
-
-        assert (
-            len(value[0]) == 2 for value in s["frequency_ranges_hz"].values()
-        )
-
-        assert (
-            isinstance(value[1], (float, int))
-            for value in s["frequency_ranges_hz"].values()
-        )
-
-    def init_KF(self, feature: str) -> None:
-        for f_band in self.s["kalman_filter_settings"]["frequency_bands"]:
-            for channel in self.ch_names:
-                self.KF_dict[
-                    "_".join([channel, feature, f_band])
-                ] = nm_kalmanfilter.define_KF(
-                    self.s["kalman_filter_settings"]["Tp"],
-                    self.s["kalman_filter_settings"]["sigma_w"],
-                    self.s["kalman_filter_settings"]["sigma_v"],
-                )
-
-    def update_KF(self, feature_calc: float, KF_name: str) -> float:
-        if KF_name in self.KF_dict:
-            self.KF_dict[KF_name].predict()
-            self.KF_dict[KF_name].update(feature_calc)
-            feature_calc = self.KF_dict[KF_name].x[0]
-        return feature_calc
-
-    def estimate_osc_features(
-        self,
-        features_compute: dict,
-        data: np.ndarray,
-        feature_name: np.ndarray,
-        est_name: str,
-    ):
-        for feature_est_name in list(self.s[est_name]["features"].keys()):
-            if self.s[est_name]["features"][feature_est_name] is True:
-                # switch case for feature_est_name
-                match feature_est_name:
-                    case "mean":
-                        features_compute[
-                            f"{feature_name}_{feature_est_name}"
-                        ] = np.nanmean(data)
-                    case "median":
-                        features_compute[
-                            f"{feature_name}_{feature_est_name}"
-                        ] = np.nanmedian(data)
-                    case "std":
-                        features_compute[
-                            f"{feature_name}_{feature_est_name}"
-                        ] = np.nanstd(data)
-                    case "max":
-                        features_compute[
-                            f"{feature_name}_{feature_est_name}"
-                        ] = np.nanmax(data)
-
-        return features_compute
-
-
-class FFT(OscillatoryFeature):
-    def __init__(
-        self,
-        settings: dict,
-        ch_names: Iterable[str],
-        sfreq: float,
-    ) -> None:
-        super().__init__(settings, ch_names, sfreq)
-
-        if self.s["fft_settings"]["log_transform"]:
-            self.log_transform = True
-        else:
-            self.log_transform = False
-
-        window_ms = self.s["fft_settings"]["windowlength_ms"]
-        self.window_samples = int(-np.floor(window_ms / 1000 * sfreq))
-        self.freqs = fft.rfftfreq(
-            -self.window_samples, 1 / np.floor(self.sfreq)
-        )
-
-        self.feature_params = []
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            for fband, f_range in self.f_ranges_dict.items():
-                idx_range = np.where(
-                    (self.freqs >= f_range[0]) & (self.freqs < f_range[1])
-                )[0]
-                feature_name = "_".join([ch_name, "fft", fband])
-                self.feature_params.append((ch_idx, feature_name, idx_range))
-
-    @staticmethod
-    def test_settings(s: dict, ch_names: Iterable[str], sfreq: int | float):
-        OscillatoryFeature.test_settings_osc(s, ch_names, sfreq, "fft_settings")
-
-    def calc_feature(self, data: np.ndarray, features_compute: dict) -> dict:
-        data = data[:, self.window_samples :]
-        Z = np.abs(fft.rfft(data))
-
-        if self.log_transform:
-            Z = np.log10(Z)
-
-        for ch_idx, feature_name, idx_range in self.feature_params:
-            Z_ch = Z[ch_idx, idx_range]
-
-            features_compute = self.estimate_osc_features(
-                features_compute, Z_ch, feature_name, "fft_settings"
-            )
-
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            if self.s["fft_settings"]["return_spectrum"]:
-                features_compute.update(
-                    {
-                        f"{ch_name}_fft_psd_{str(f)}": Z[ch_idx][idx]
-                        for idx, f in enumerate(self.freqs.astype(int))
-                    }
-                )
-
-        return features_compute
-
-
-class Welch(OscillatoryFeature):
-    def __init__(
-        self,
-        settings: dict,
-        ch_names: Iterable[str],
-        sfreq: float,
-    ) -> None:
-        super().__init__(settings, ch_names, sfreq)
-
-        self.log_transform = self.s["welch_settings"]["log_transform"]
-
-        self.feature_params = []
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            for fband, f_range in self.f_ranges_dict.items():
-                feature_name = "_".join([ch_name, "welch", fband])
-                self.feature_params.append((ch_idx, feature_name, f_range))
-
-    @staticmethod
-    def test_settings(s: dict, ch_names: Iterable[str], sfreq: int | float):
-        OscillatoryFeature.test_settings_osc(
-            s, ch_names, sfreq, "welch_settings"
-        )
-
-    def calc_feature(self, data: np.ndarray, features_compute: dict) -> dict:
-        freqs, Z = signal.welch(
-            data,
-            fs=self.sfreq,
-            window="hann",
-            nperseg=self.sfreq,
-            noverlap=None,
-        )
-
-        if self.log_transform:
-            Z = np.log10(Z)
-
-        for ch_idx, feature_name, f_range in self.feature_params:
-            Z_ch = Z[ch_idx]
-
-            idx_range = np.where((freqs >= f_range[0]) & (freqs <= f_range[1]))[
-                0
-            ]
-
-            features_compute = self.estimate_osc_features(
-                features_compute,
-                Z_ch[idx_range],
-                feature_name,
-                "welch_settings",
-            )
-
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            if self.s["welch_settings"]["return_spectrum"]:
-                features_compute.update(
-                    {
-                        f"{ch_name}_welch_psd_{str(f)}": Z[ch_idx][idx]
-                        for idx, f in enumerate(freqs.astype(int))
-                    }
-                )
-
-        return features_compute
-
-
-class STFT(OscillatoryFeature):
-    def __init__(
-        self,
-        settings: dict,
-        ch_names: Iterable[str],
-        sfreq: float,
-    ) -> None:
-        super().__init__(settings, ch_names, sfreq)
-
-        self.nperseg = int(self.s["stft_settings"]["windowlength_ms"])
-        self.log_transform = self.s["stft_settings"]["log_transform"]
-
-        self.feature_params = []
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            for fband, f_range in self.f_ranges_dict.items():
-                feature_name = "_".join([ch_name, "stft", fband])
-                self.feature_params.append((ch_idx, feature_name, f_range))
-
-    @staticmethod
-    def test_settings(s: dict, ch_names: Iterable[str], sfreq: int | float):
-        OscillatoryFeature.test_settings_osc(
-            s, ch_names, sfreq, "stft_settings"
-        )
-
-    def calc_feature(self, data: np.ndarray, features_compute: dict) -> dict:
-        freqs, _, Zxx = signal.stft(
-            data,
-            fs=self.sfreq,
-            window="hamming",
-            nperseg=self.nperseg,
-            boundary="even",
-        )
-        Z = np.abs(Zxx)
-        if self.log_transform:
-            Z = np.log10(Z)
-        for ch_idx, feature_name, f_range in self.feature_params:
-            Z_ch = Z[ch_idx]
-            idx_range = np.where((freqs >= f_range[0]) & (freqs <= f_range[1]))[
-                0
-            ]
-
-            features_compute = self.estimate_osc_features(
-                features_compute,
-                Z_ch[idx_range, :],
-                feature_name,
-                "stft_settings",
-            )
-
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            if self.s["stft_settings"]["return_spectrum"]:
-                Z_ch_mean = Z[ch_idx].mean(axis=1)
-                features_compute.update(
-                    {
-                        f"{ch_name}_stft_psd_{str(f)}": Z_ch_mean[idx]
-                        for idx, f in enumerate(freqs.astype(int))
-                    }
-                )
-
-        return features_compute
-
-
-class BandPower(OscillatoryFeature):
-    def __init__(
-        self,
-        settings: dict,
-        ch_names: Iterable[str],
-        sfreq: float,
-        use_kf: bool = None,
-    ) -> None:
-        super().__init__(settings, ch_names, sfreq)
-        bp_settings = self.s["bandpass_filter_settings"]
-
-        self.bandpass_filter = nm_filter.MNEFilter(
-            f_ranges=list(self.f_ranges_dict.values()),
-            sfreq=self.sfreq,
-            filter_length=self.sfreq - 1,
-            verbose=False,
-        )
-
-        self.log_transform = bp_settings["log_transform"]
-
-        if use_kf is True or (
-            use_kf is None and bp_settings["kalman_filter"] is True
-        ):
-            self.init_KF("bandpass_activity")
-
-        bp_features = ["activity", "mobility", "complexity"]
-        seglengths = bp_settings["segment_lengths_ms"]
-
-        self.feature_params = []
-        for ch_idx, ch_name in enumerate(self.ch_names):
-            for f_band_idx, f_band in enumerate(self.f_ranges_dict):
-                seglength_ms = seglengths[f_band]
-                seglen = int(np.floor(self.sfreq / 1000 * seglength_ms))
-                for bp_feature, v in bp_settings["bandpower_features"].items():
-                    if v is True:
-                        if bp_feature not in bp_features:
-                            raise ValueError()
-                        feature_name = "_".join(
-                            [ch_name, "bandpass", bp_feature, f_band]
-                        )
-                        self.feature_params.append(
-                            (
-                                ch_idx,
-                                ch_name,
-                                f_band,
-                                f_band_idx,
-                                seglen,
-                                bp_feature,
-                                feature_name,
-                            )
-                        )
-
-    @staticmethod
-    def test_settings(s: dict, ch_names: Iterable[str], sfreq: int | float):
-        OscillatoryFeature.test_settings_osc(
-            s, ch_names, sfreq, "bandpass_filter_settings"
-        )
-
-        assert (
-            isinstance(value, bool)
-            for value in s["bandpass_filter_settings"][
-                "bandpower_features"
-            ].values()
-        )
-
-        assert any(
-            value is True
-            for value in s["bandpass_filter_settings"][
-                "bandpower_features"
-            ].values()
-        ), "Set at least one bandpower_feature to True."
-
-        for fband_name, seg_length_fband in s["bandpass_filter_settings"][
-            "segment_lengths_ms"
-        ].items():
-            assert isinstance(seg_length_fband, int), (
-                f"bandpass segment_lengths_ms for {fband_name} "
-                f"needs to be of type int, got {seg_length_fband}"
-            )
-
-            assert seg_length_fband <= s["segment_length_features_ms"], (
-                f"segment length {seg_length_fband} needs to be smaller than "
-                f" s['segment_length_features_ms'] = {s['segment_length_features_ms']}"
-            )
-
-        for fband_name in list(s["frequency_ranges_hz"].keys()):
-            assert fband_name in list(
-                s["bandpass_filter_settings"]["segment_lengths_ms"].keys()
-            ), (
-                f"frequency range {fband_name} "
-                "needs to be defined in s['bandpass_filter_settings']['segment_lengths_ms']"
-            )
-
-    def calc_feature(self, data: np.ndarray, features_compute: dict) -> dict:
-        data = self.bandpass_filter.filter_data(data)
-
-        for (
-            ch_idx,
-            ch_name,
-            f_band,
-            f_band_idx,
-            seglen,
-            bp_feature,
-            feature_name,
-        ) in self.feature_params:
-            if bp_feature == "activity":
-                if self.log_transform:
-                    feature_calc = np.log10(
-                        np.var(data[ch_idx, f_band_idx, -seglen:])
-                    )
-                else:
-                    feature_calc = np.var(data[ch_idx, f_band_idx, -seglen:])
-            elif bp_feature == "mobility":
-                deriv_variance = np.var(
-                    np.diff(data[ch_idx, f_band_idx, -seglen:])
-                )
-                feature_calc = np.sqrt(
-                    deriv_variance / np.var(data[ch_idx, f_band_idx, -seglen:])
-                )
-            elif bp_feature == "complexity":
-                dat_deriv = np.diff(data[ch_idx, f_band_idx, -seglen:])
-                deriv_variance = np.var(dat_deriv)
-                mobility = np.sqrt(
-                    deriv_variance / np.var(data[ch_idx, f_band_idx, -seglen:])
-                )
-                dat_deriv_2 = np.diff(dat_deriv)
-                dat_deriv_2_var = np.var(dat_deriv_2)
-                deriv_mobility = np.sqrt(dat_deriv_2_var / deriv_variance)
-                feature_calc = deriv_mobility / mobility
-
-            if self.KF_dict and (bp_feature == "activity"):
-                feature_calc = self.update_KF(feature_calc, feature_name)
-
-            features_compute[feature_name] = np.nan_to_num(feature_calc)
-
-        return features_compute
+from collections.abc import Iterable
+import numpy as np
+from itertools import product
+
+from py_neuromodulation.nm_types import NMBaseModel
+from pydantic import field_validator
+from typing import TYPE_CHECKING
+
+from py_neuromodulation.nm_features import NMFeature
+from py_neuromodulation.nm_types import BoolSelector
+
+if TYPE_CHECKING:
+    from py_neuromodulation.nm_settings import NMSettings
+    from py_neuromodulation.nm_kalmanfilter import KalmanSettings
+
+
+class OscillatoryFeatures(BoolSelector):
+    mean: bool = True
+    median: bool = False
+    std: bool = False
+    max: bool = False
+
+
+class OscillatorySettings(NMBaseModel):
+    windowlength_ms: int = 1000
+    log_transform: bool = True
+    features: OscillatoryFeatures = OscillatoryFeatures(
+        mean=True, median=False, std=False, max=False
+    )
+    return_spectrum: bool = False
+
+
+ESTIMATOR_DICT = {
+    "mean": np.nanmean,
+    "median": np.nanmedian,
+    "std": np.nanstd,
+    "max": np.nanmax,
+}
+
+
+class OscillatoryFeature(NMFeature):
+    def __init__(
+        self, settings: "NMSettings", ch_names: Iterable[str], sfreq: int
+    ) -> None:
+        settings.validate()
+        self.settings: OscillatorySettings  # Assignment in subclass __init__
+        self.osc_feature_name: str  # Required for output
+
+        self.sfreq = int(sfreq)
+        self.ch_names = ch_names
+
+        self.frequency_ranges = settings.frequency_ranges_hz
+
+        # Test settings
+        assert self.settings.windowlength_ms <= settings.segment_length_features_ms, (
+            f"oscillatory feature windowlength_ms = ({self.settings.windowlength_ms})"
+            f"needs to be smaller than"
+            f"settings['segment_length_features_ms'] = {settings.segment_length_features_ms}",
+        )
+
+
+class FFT(OscillatoryFeature):
+    def __init__(
+        self,
+        settings: "NMSettings",
+        ch_names: Iterable[str],
+        sfreq: int,
+    ) -> None:
+        from scipy.fft import rfftfreq
+
+        self.osc_feature_name = "fft"
+        self.settings = settings.fft_settings
+        # super.__init__ needs osc_feature_name and settings
+        super().__init__(settings, ch_names, sfreq)
+
+        window_ms = self.settings.windowlength_ms
+
+        self.window_samples = int(-np.floor(window_ms / 1000 * sfreq))
+        self.freqs = rfftfreq(-self.window_samples, 1 / np.floor(self.sfreq))
+
+        # Pre-calculate frequency ranges
+        self.idx_range = [
+            (
+                f_band,
+                np.where((self.freqs >= f_range[0]) & (self.freqs < f_range[1]))[0],
+            )
+            for f_band, f_range in self.frequency_ranges.items()
+        ]
+
+        self.estimators = [
+            (est, ESTIMATOR_DICT[est]) for est in self.settings.features.get_enabled()
+        ]
+
+    def calc_feature(self, data: np.ndarray) -> dict:
+        data = data[:, self.window_samples :]
+
+        from scipy.fft import rfft
+
+        Z = np.abs(rfft(data))  # type: ignore
+
+        if self.settings.log_transform:
+            Z = np.log10(Z)
+
+        feature_results = {}
+
+        for f_band_name, idx_range in self.idx_range:
+            # TODO Can we get rid of this for-loop? Hard to vectorize windows of different lengths...
+            Z_band = Z[:, idx_range]  # Data for all channels
+
+            for est_name, est_fun in self.estimators:
+                result = est_fun(Z_band, axis=1)
+
+                for ch_idx, ch_name in enumerate(self.ch_names):
+                    feature_results[
+                        f"{ch_name}_{self.osc_feature_name}_{f_band_name}_{est_name}"
+                    ] = result[ch_idx]
+
+        if self.settings.return_spectrum:
+            combinations = product(enumerate(self.ch_names), enumerate(self.freqs))
+            for (ch_idx, ch_name), (idx, f) in combinations:
+                feature_results[f"{ch_name}_fft_psd_{int(f)}"] = Z[ch_idx][idx]
+
+        return feature_results
+
+
+class Welch(OscillatoryFeature):
+    def __init__(
+        self,
+        settings: "NMSettings",
+        ch_names: Iterable[str],
+        sfreq: int,
+    ) -> None:
+        from scipy.fft import rfftfreq
+
+        self.osc_feature_name = "welch"
+        self.settings = settings.welch_settings
+        # super.__init__ needs osc_feature_name and settings
+        super().__init__(settings, ch_names, sfreq)
+
+        self.freqs = rfftfreq(self.sfreq, 1 / self.sfreq)
+
+        self.idx_range = [
+            (
+                f_band,
+                np.where((self.freqs >= f_range[0]) & (self.freqs < f_range[1]))[0],
+            )
+            for f_band, f_range in self.frequency_ranges.items()
+        ]
+
+        self.estimators = [
+            (est, ESTIMATOR_DICT[est]) for est in self.settings.features.get_enabled()
+        ]
+
+    def calc_feature(self, data: np.ndarray) -> dict:
+        from scipy.signal import welch
+
+        _, Z = welch(
+            data,
+            fs=self.sfreq,
+            window="hann",
+            nperseg=self.sfreq,
+            noverlap=None,
+        )
+
+        if self.settings.log_transform:
+            Z = np.log10(Z)
+
+        feature_results = {}
+
+        for f_band_name, idx_range in self.idx_range:
+            Z_band = Z[:, idx_range]
+
+            for est_name, est_fun in self.estimators:
+                result = est_fun(Z_band, axis=1)
+
+                for ch_idx, ch_name in enumerate(self.ch_names):
+                    feature_results[
+                        f"{ch_name}_{self.osc_feature_name}_{f_band_name}_{est_name}"
+                    ] = result[ch_idx]
+
+        if self.settings.return_spectrum:
+            combinations = product(enumerate(self.ch_names), enumerate(self.freqs))
+            for (ch_idx, ch_name), (idx, f) in combinations:
+                feature_results[f"{ch_name}_welch_psd_{str(f)}"] = Z[ch_idx][idx]
+
+        return feature_results
+
+
+class STFT(OscillatoryFeature):
+    def __init__(
+        self,
+        settings: "NMSettings",
+        ch_names: Iterable[str],
+        sfreq: int,
+    ) -> None:
+        from scipy.fft import rfftfreq
+
+        self.osc_feature_name = "stft"
+        self.settings = settings.stft_settings
+        # super.__init__ needs osc_feature_name and settings
+        super().__init__(settings, ch_names, sfreq)
+
+        self.nperseg = self.settings.windowlength_ms
+
+        self.freqs = rfftfreq(self.nperseg, 1 / self.sfreq)
+
+        self.idx_range = [
+            (
+                f_band,
+                np.where((self.freqs >= f_range[0]) & (self.freqs <= f_range[1]))[0],
+            )
+            for f_band, f_range in self.frequency_ranges.items()
+        ]
+
+        self.estimators = [
+            (est, ESTIMATOR_DICT[est]) for est in self.settings.features.get_enabled()
+        ]
+
+    def calc_feature(self, data: np.ndarray) -> dict:
+        from scipy.signal import stft
+
+        _, _, Zxx = stft(
+            data,
+            fs=self.sfreq,
+            window="hamming",
+            nperseg=self.nperseg,
+            boundary="even",
+        )
+
+        Z = np.abs(Zxx)
+        if self.settings.log_transform:
+            Z = np.log10(Z)
+
+        feature_results = {}
+
+        for f_band_name, idx_range in self.idx_range:
+            Z_band = Z[:, idx_range, :]
+
+            for est_name, est_fun in self.estimators:
+                result = est_fun(Z_band, axis=(1, 2))
+
+                for ch_idx, ch_name in enumerate(self.ch_names):
+                    feature_results[
+                        f"{ch_name}_{self.osc_feature_name}_{f_band_name}_{est_name}"
+                    ] = result[ch_idx]
+
+        if self.settings.return_spectrum:
+            combinations = product(enumerate(self.ch_names), enumerate(self.freqs))
+            for (ch_idx, ch_name), (idx, f) in combinations:
+                feature_results[f"{ch_name}_stft_psd_{str(f)}"] = Z[ch_idx].mean(
+                    axis=1
+                )[idx]
+
+        return feature_results
+
+
+class BandpowerFeatures(BoolSelector):
+    activity: bool = True
+    mobility: bool = False
+    complexity: bool = False
+
+
+###################################
+######## BANDPOWER FEATURE ########
+###################################
+
+
+class BandpassSettings(NMBaseModel):
+    segment_lengths_ms: dict[str, int] = {
+        "theta": 1000,
+        "alpha": 500,
+        "low_beta": 333,
+        "high_beta": 333,
+        "low_gamma": 100,
+        "high_gamma": 100,
+        "HFA": 100,
+    }
+    bandpower_features: BandpowerFeatures = BandpowerFeatures()
+    log_transform: bool = True
+    kalman_filter: bool = False
+
+    @field_validator("segment_lengths_ms")
+    @classmethod
+    # Replace spaces with underscores in frequency band names
+    def fbands_spaces_to_underscores(cls, segment_lengths_ms: dict[str, int]):
+        return {k.replace(" ", "_"): v for k, v in segment_lengths_ms.items()}
+
+    @field_validator("bandpower_features")
+    @classmethod
+    def bandpower_features_validator(cls, bandpower_features: BandpowerFeatures):
+        assert (
+            len(bandpower_features.get_enabled()) > 0
+        ), "Set at least one bandpower_feature to True."
+
+        return bandpower_features
+
+    def validate_fbands(self, settings: "NMSettings") -> None:
+        # Ensure that each freq-band is defined in the global settings
+        for fband_name in settings.frequency_ranges_hz.keys():
+            assert fband_name in self.segment_lengths_ms, (
+                f"frequency range {fband_name} "
+                "needs to be defined in settings.bandpass_filter_settings.segment_lengths_ms]"
+            )
+
+        # Ensure that segment length for each freq-band is smaller than the feature segment length setting
+        for fband_name, seg_length_fband in self.segment_lengths_ms.items():
+            assert seg_length_fband <= settings.segment_length_features_ms, (
+                f"segment length {seg_length_fband} needs to be smaller than "
+                f" settings['segment_length_features_ms'] = {settings.segment_length_features_ms}"
+            )
+
+
+class BandPower(NMFeature):
+    def __init__(
+        self,
+        settings: "NMSettings",
+        ch_names: Iterable[str],
+        sfreq: float,
+        use_kf: bool | None = None,
+    ) -> None:
+        settings.validate()
+
+        self.bp_settings: BandpassSettings = settings.bandpass_filter_settings
+        self.kalman_filter_settings: KalmanSettings = settings.kalman_filter_settings
+        self.sfreq = sfreq
+        self.ch_names = ch_names
+        self.KF_dict: dict = {}
+
+        from py_neuromodulation.nm_filter import MNEFilter
+
+        self.bandpass_filter = MNEFilter(
+            f_ranges=[
+                tuple(frange) for frange in settings.frequency_ranges_hz.values()
+            ],
+            sfreq=self.sfreq,
+            filter_length=self.sfreq - 1,
+            verbose=False,
+        )
+
+        if use_kf or (use_kf is None and self.bp_settings.kalman_filter):
+            self.init_KF("bandpass_activity")
+
+        seglengths = self.bp_settings.segment_lengths_ms
+
+        self.feature_params = []
+        for ch_idx, ch_name in enumerate(self.ch_names):
+            for f_band_idx, f_band in enumerate(settings.frequency_ranges_hz.keys()):
+                seglength_ms = seglengths[f_band]
+                seglen = int(np.floor(self.sfreq / 1000 * seglength_ms))
+                for bp_feature in self.bp_settings.bandpower_features.get_enabled():
+                    feature_name = "_".join([ch_name, "bandpass", bp_feature, f_band])
+                    self.feature_params.append(
+                        (
+                            ch_idx,
+                            f_band_idx,
+                            seglen,
+                            bp_feature,
+                            feature_name,
+                        )
+                    )
+
+    def init_KF(self, feature: str) -> None:
+        from py_neuromodulation.nm_kalmanfilter import define_KF
+
+        for f_band in self.kalman_filter_settings.frequency_bands:
+            for channel in self.ch_names:
+                self.KF_dict["_".join([channel, feature, f_band])] = define_KF(
+                    self.kalman_filter_settings.Tp,
+                    self.kalman_filter_settings.sigma_w,
+                    self.kalman_filter_settings.sigma_v,
+                )
+
+    def update_KF(self, feature_calc: np.floating, KF_name: str) -> np.floating:
+        if KF_name in self.KF_dict:
+            self.KF_dict[KF_name].predict()
+            self.KF_dict[KF_name].update(feature_calc)
+            feature_calc = self.KF_dict[KF_name].x[0]
+        return feature_calc
+
+    def calc_feature(self, data: np.ndarray) -> dict:
+        data = self.bandpass_filter.filter_data(data)
+
+        feature_results = {}
+
+        for (
+            ch_idx,
+            f_band_idx,
+            seglen,
+            bp_feature,
+            feature_name,
+        ) in self.feature_params:
+            feature_results[feature_name] = self.calc_bp_feature(
+                bp_feature, feature_name, data[ch_idx, f_band_idx, -seglen:]
+            )
+
+        return feature_results
+
+    def calc_bp_feature(self, bp_feature, feature_name, data):
+        match bp_feature:
+            case "activity":
+                feature_calc = np.var(data)
+                if self.bp_settings.log_transform:
+                    feature_calc = np.log10(feature_calc)
+                if self.KF_dict:
+                    feature_calc = self.update_KF(feature_calc, feature_name)
+            case "mobility":
+                feature_calc = np.sqrt(np.var(np.diff(data)) / np.var(data))
+            case "complexity":
+                feature_calc = self.calc_complexity(data)
+            case _:
+                raise ValueError(f"Unknown bandpower feature: {bp_feature}")
+
+        return np.nan_to_num(feature_calc)
+
+    @staticmethod
+    def calc_complexity(data: np.ndarray) -> float:
+        dat_deriv = np.diff(data)
+        deriv_variance = np.var(dat_deriv)
+        mobility = np.sqrt(deriv_variance / np.var(data))
+        dat_deriv_2_var = np.var(np.diff(dat_deriv))
+        deriv_mobility = np.sqrt(dat_deriv_2_var / deriv_variance)
+
+        return deriv_mobility / mobility