Electrophysiological, imaging, behavior and synchronization data of Cerebellar activity is triggered by reach endpoint during learning of a complex locomotor task

LocoReach Dataset — Raw Electrophysiology and Behavior Recordings Overview This dataset contains raw in vivo extracellular recordings from the cerebellum(molecular layer interneurons, MLI, and Purkinje cells, PC) of head-fixed micewalking on a motorized, runged treadmill (LocoReach). Recordings were performed while animals walked on the wheel with regular (equally-spaced) rungs requiring precise step adjustments. Each trial lasts approximately 60 seconds and contains: Single-unit electrophysiology (MLI or PC) Behavioral video (side and bottom view of the animal on the ladder) Rotary encoder (treadmill speed) DAQ synchronization signals Animals and Recordings | Animal ID | Sex | MLI cells | MLI trials | PC cells | PC trials | Recording days ||--------------|-----|-----------|------------|----------|-----------|----------------|| 220211_f38 | F | 6 | 27 | 3 | 13 | 8 || 220214_f43 | F | 3 | 7 | 6 | 19 | 8 || 220205_f57 | F | 4 | 20 | 3 | 15 | 6 || 220205_f61 | F | 10 | 36 | 2 | 7 | 7 || 220507_m81 | M | 8 | 27 | 3 | 13 | 8 || 220507_m90 | M | 8 | 40 | 2 | 6 | 8 || 220525_m19 | M | 5 | 18 | 4 | 20 | 8 || 220525_m27 | M | 6 | 30 | 5 | 21 | 10 || 220525_m28 | M | 7 | 34 | 4 | 19 | 10 || 220716_f65 | F | 4 | 17 | 1 | 5 | 5 || 220716_f67 | F | 3 | 12 | 1 | 5 | 4 || **Total** | | **64** | **268** | **34** | **143** | **—** | Animal ID encodes implantation date (YYMMDD) and animal number.Total: 410 trial folders across 11 animals. Dataset Structure The dataset is distributed as one .tar.gz archive per animal: 220211_f38.tar.gz220214_f43.tar.gz...220716_f67.tar.gz Each archive expands to: <ANIMAL_ID>/ <YYYY.MM.DD_NNN>/ # recording session folder locomotionEphys2Motor60sec_TTT/ # one folder per trial (TTT = trial number) AxoPatch200_2.ma # electrophysiology patch-clamp amplifier signal DaqDevice.ma # DAQ synchronization and analog channels RotaryEncoder.ma # rotary encoder (running speed) CameraGigEBehavior/ video_000.avi # behavioral video (MJPG, converted from .ma) daqResult.ma # camera trigger / frame timing File Formats All .ma files are HDF5 files recorded by ACQ4 (open-source data acquisition software,https://github.com/acq4/acq4). They can be opened with any HDF5 reader (e.g. h5py inPython, HDFView, MATLAB's h5read). Electrophysiology files (AxoPatch200_2.ma, DaqDevice.ma, RotaryEncoder.ma) Key HDF5 datasets:- /data — 1-D array of samples (float32 or int16, depending on channel)- /info/0/ — channel metadata including sampling rate (rate) and physical units (units) Camera trigger file (CameraGigEBehavior/daqResult.ma) - /data — 1-D array of the camera exposure signal (TTL pulses), same format as above Behavioral video (CameraGigEBehavior/video_000.avi) Videos were originally recorded as ACQ4 .ma (HDF5) files and converted to MJPG.avi for broad compatibility. Each frame is an 8-bit grayscale image. Frametimestamps are embedded in the original .ma file (/info/0/values). Videos were recorded at 200 fps. To read frames in Python:#pythonimport cv2cap = cv2.VideoCapture('video_000.avi')while True: ret, frame = cap.read() if not ret: break # frame is (H, W, 3) uint8 BGR; convert with cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)cap.release() Acquisition Details Electrophysiology: patch-clamp amplifier (Axopatch 200B), loose-seal cell-attached configuration; signals digitized at 20 kHz Camera: GigE machine-vision camera, 200 fps, bottom and side view Locomotion apparatus: motorized rung ladder (LocoReach); regular rung spacing Related Resources Analysis code: https://github.com/mgraupe/LocoRungs ACQ4 documentation: https://acq4.readthedocs.io