tetrode_analyses

Loading helpers for Open Ephys tetrode acquisitions.

Public API (tetrode_analyses / ta)

Function	Purpose
print_session_tree(acq_dir)	Print record nodes, experiments, recordings, and continuous streams (with channel/sample counts) for an acquisition.
open_tetrode_dataarray(acq_dir, ...)	Load one continuous stream as a tetrode-shaped lazy xarray.DataArray (2d_flat_index, 2d_multi_index, or 3d layout).
load_channel_map(source, *, block_index=None)	Load the tetrode channel map (oe_indices, enabled, tt2ch_ixs) from a settings.xml path, an acquisition / record-node directory, or a SpikeInterface OpenEphysBinaryRecordingExtractor.
build_tetrode_probegroup(n_tetrodes, ...)	Build a probeinterface.ProbeGroup of generic tetrodes (identity-wired).
attach_tetrode_probegroup(recording, ...)	Return a SpikeInterface recording reordered into channel-map order and grouped by tetrode (group + tetrode properties, plus a ProbeGroup when geometry=True).

The channel map / tetrode grouping is an experiment-level property (read from settings.xml's Channel Map processor), while channel names are a per-recording property (each recording's structure.oebin).

Give a SpikeInterface extractor the tetrode grouping

A raw OpenEphysBinaryRecordingExtractor is a flat list of channels with no probe and no grouping. attach_tetrode_probegroup applies the same channel map open_tetrode_dataarray uses, so the recording can be split / preprocessed / sorted per tetrode (cf. SpikeInterface's Work with tetrodes and Process a recording by channel group how-tos):

import tetrode_analyses as ta
from spikeinterface.extractors.extractor_classes import (
    OpenEphysBinaryRecordingExtractor,
)

extractor = OpenEphysBinaryRecordingExtractor(ACQ_DIR, stream_id="0")
grouped = ta.attach_tetrode_probegroup(extractor)

grouped.get_channel_groups()      # [0, 0, 0, 0, 1, 1, 1, 1, ...]
by_tetrode = grouped.split_by("group")   # one recording per tetrode
by_tetrode[0].get_channel_ids()   # ['CH40', 'CH38', 'CH36', 'CH34'] == TT1

attach_tetrode_probegroup / build_tetrode_probegroup require probeinterface, declared in the optional si extra (pip install -e '.[si]' / uv sync --extra si). The SpikeInterface extractor is supplied by the caller, so spikeinterface itself is not a dependency of this package.

Analysis library modules

Heavier submodules are imported directly (not re-exported from __init__, to keep import tetrode_analyses light). They require the optional analysis extra (uv sync --extra analysis: ecephys, matplotlib, seaborn, zarr).

Module	Key functions
tetrode_analyses.lfp	make_lfp, open_lfps_dataarray — produce / load the 625 Hz LFP Zarr (sorting extra).
tetrode_analyses.experiment	ExperimentParams, load/save_experiment_params, read_session_t0_unix, dt2t/t2dt, get_light_dark_periods, get_deprivation_period — WNE-style experiment_params.json and wall-clock↔session-time mapping.
tetrode_analyses.power	extract_instantaneous_bandpower (bandpass + Hilbert), compute_stft_bandpowers (4 s DPSS STFT), both averaged per tetrode; save/open helpers; replace_outliers / replace_outliers_per_tetrode (histogram-gap artifact removal). BANDS defines the frequency bands.
tetrode_analyses.plotting	plot_swa_timetrace, plot_lights_overlay, plot_deprivation_overlay, plot_swa_overview, plot_swa_small_multiples — SWA timetraces with light/dark + deprivation overlays.

See analyses/swa/ for the end-to-end SWA (delta power) pipeline built on these.