Manually setting SpikeGLX recording sampling frequency

[louiskang]

I'd like to manually set the sampling frequency of a Neuropixels recording loaded via read_spikeglx. The reason is that my analysis requires precise synchronization of traces, and environmental fluctuations (e.g. temperature) can create slight discrepancies from the calibrated sampling frequency on an experiment-by-experiment basis. Using the sync channel, I can compute an empirical sampling frequency for each experiment, and I would like to use these values. Is there a way to do this? If not, can the simple feature of manually setting recording properties be introduced? My current workaround is directly editing the .meta file. Thanks!

[louiskang]

I wrote this subclass, which also seems to work for .get_times() because .time_vector is None for SpikeGLX imports. Feel free to close this issue if it's not a priority to override recording properties in general.

import spikeinterface.full as si

class OverriddenFrequencyRecording(si.BaseRecording):
    """
    Class to override the sampling frequency of a recording.

    Parameters
    ----------
    parent_recording: BaseRecording
    sampling_frequency: float
    """

    def __init__(self, parent_recording, sampling_frequency):
        si.BaseRecording.__init__(
            self,
            sampling_frequency=sampling_frequency,
            channel_ids=parent_recording.get_channel_ids(),
            dtype=parent_recording.get_dtype(),
        )

        # link recording segment
        for parent_segment in parent_recording._recording_segments:
            sub_segment = OverriddenFrequencyRecordingSegment(parent_segment, sampling_frequency)
            self.add_recording_segment(sub_segment)

        # copy annotation and properties
        parent_recording.copy_metadata(self)
        self._parent = parent_recording

        # update dump dict
        self._kwargs = {
            "parent_recording": parent_recording,
            "sampling_frequency": sampling_frequency,
        }

class OverriddenFrequencyRecordingSegment(si.BaseRecordingSegment):
    """
    Class to override the sampling frequency of a recording segment.
    """

    def __init__(self, parent_recording_segment, sampling_frequency):
        d = parent_recording_segment.get_times_kwargs()
        d = d.copy()
        d["sampling_frequency"] = sampling_frequency
        si.BaseRecordingSegment.__init__(self, **d)
        self._parent_recording_segment = parent_recording_segment

    def get_num_samples(self) -> int:
        return self._parent_recording_segment.get_num_samples()

    def get_traces(
        self,
        start_frame: int | None = None,
        end_frame: int | None = None,
        channel_indices: list | None = None,
    ) -> np.ndarray:
        return self._parent_recording_segment.get_traces(start_frame, end_frame, channel_indices)

[h-mayorquin]

Something to discuss, maybe we should allow override of any recording sampling_frequency with a special method?

[alejoe91]

Hi @louiskang

You can also set_times to a recording with the synchronized timestamps. Would that work? I think it would be more precise then setting a "global" fixed sampling rate, as indeed you might have fluctuations over the course of the experiment

[louiskang]

That’s true, but in my case I downsample the data before LFP processing, and I believe that ResampleRecording just looks for sampling_frequency. Thanks!

On Wed, Dec 10, 2025 at 18:34 Alessio Buccino @.***> wrote:

alejoe91 left a comment (SpikeInterface/spikeinterface#4248) https://github.com/SpikeInterface/spikeinterface/issues/4248#issuecomment-3636174739

Hi @louiskang https://github.com/louiskang

You can also set_times to a recording with the synchronized timestamps. Would that work? I think it would be more precise then setting a "global" fixed sampling rate, as indeed you might have fluctuations over the course of the experiment

— Reply to this email directly, view it on GitHub https://github.com/SpikeInterface/spikeinterface/issues/4248#issuecomment-3636174739, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGN4FPZZ37CBP7RVGZWFU3L4A7SL7AVCNFSM6AAAAACOI73YISVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMZWGE3TINZTHE . You are receiving this because you were mentioned.Message ID: @.***>