Meeting Notes
- Date: 2025-08-26
- Time: 09:00 (PT)
- Location: Teams Meeting
- Presentations: @jeromelecoq @severine2305 @maierav
Agenda
- Discussion on motor-oddball regimen design. @jeromelecoq
- Follow-up on recording set of areas with Neuropixels. Discussion of pilot. @severine2305
- Update on analysis of SLAP2 data. @maierav
Meeting Recording
Meeting Notes
Experimental Design and Stimulus Protocols: Jerome led a detailed review of the experimental design for upcoming neuropixels and two-photon experiments, outlining session types, stimulus blocks, and control protocols, with input from Alexander, Karim, and Farzaneh on technical and scientific aspects.
Session Structure Overview: Jerome described the planned session types, including standard mismatch, sensory motor mismatch, sequence mismatch, and duration mismatch, with each session containing one oddball block and several control blocks. The design includes both two-photon and neuropixels cohorts, with mice trained in either sensory motor or sequential contexts before transitioning to recording rigs for context-specific recordings.
Stimulus Block Details: Jerome explained the rationale and timing for each stimulus block, such as orientation tuning at session onset and after oddball blocks, standard mismatch blocks with randomized alternative orientations, and control blocks for contiguous drifting gratings and jittered durations. Adjustments were made to the number of repeats and block durations to optimize session length and data quality.
Stimulus Table Generation and Randomization: Jerome outlined the process for generating and randomizing stimulus tables using Python scripts, which are then executed by Bonsai. Alexander raised concerns about true versus pseudo-randomization to avoid statistical artifacts, and Jerome agreed to validate the randomization and ensure accurate stimulus-event alignment.
Validation and Quality Control: Jerome initiated validation of the stimulus tables by plotting parameter sweeps and block content, inviting team members to review the tables and code for accuracy. Alexander and Severine discussed the importance of double-checking event alignment and avoiding frame shifts, with plans for further collaborative validation.
Neuropixels Target Area Selection: Severine, Jerome, Karim, Farzaneh, and Alexander discussed and debated the selection of brain areas for neuropixels probe implantation, considering scientific priorities, technical constraints, and cross-lab alignment, ultimately moving toward consensus on including Vis A over RL/LGN.
Initial Area Proposal: Severine presented the agreed-upon implant targets: V1, RL, LM, MOP, ACA, and MOS, explaining the rationale for each and the technical limitations of using four-shank probes, which led to the decision to use 1.0 probes for each hole.
Scientific Rationale and Cross-Lab Considerations: Karim and Farzaneh questioned the inclusion of RL and LM, suggesting Vis A and PPC as alternatives to better capture visual-motor and multisensory processing. The discussion included considerations of previous recording success, overlap with other labs' interests, and the importance of LGN.
Consensus and Next Steps: After weighing the pros and cons, the group leaned toward replacing RL/LGN with Vis A, acknowledging technical risks and the need for rapid decisions due to implant scheduling. Severine agreed to review probe trajectories and share details for final input, with plans to involve Sara in the decision.
Pilot Timeline and Training Workflow: Jerome outlined the timeline for the neuropixels pilot, detailing mouse implantation, health checks, intrinsic imaging, training, and transition to recording, with feedback from the team on the workflow and scientific rationale.
Pilot Schedule and Milestones: Jerome presented a timeline starting with mouse implantation in early September, followed by health checks, intrinsic imaging to align implants, and progressive head-fixation training. Mice will be habituated to the neuropixels rigs before actual recordings begin in October.
Training Protocols: The training cluster involves gradually increasing head-fixation durations and exposure to the stimulus design, initially without oddballs, to acclimate mice to the experimental setup and ensure readiness for recording sessions.
Transition to Recording and Session Structure: After training, mice will undergo four recording sessions on neuropixels rigs, starting with sensory motor mismatch and including all control conditions. The workflow is designed to test the full experimental pipeline and gather feedback for refinement.
SLAP2 Data Analysis and Code Updates: Alexander and Carter provided updates on SLAP2 data analysis, highlighting code changes, timing corrections, and data packaging improvements, with Jerome and others discussing ongoing validation and future uploads.
Analysis Workflow and Code Adjustments: Alexander demonstrated the Google Colab workflow for SLAP2 data analysis, noting changes in variable names and the need for careful alignment of stimulus events with neural data. Issues with timing and data drift were identified and partially resolved.
Data Packaging and Bug Fixes: Carter reported a bug in the Bonsai script causing misaligned trial pulses, which was corrected in the latest data files. Users were advised to download updated versions for accurate analysis, and additional sessions are being uploaded despite motion artifacts in some data.
Collaborative Validation and Next Steps: The team agreed to continue validating the analysis pipeline, with Alexander inviting others to review and expand the code, and Carter ensuring ongoing data uploads and documentation of known issues.
Consideration of Zebra Stimulus Protocol: Jerome introduced the Zebra stimulus protocol from Carandini and Arry's lab, suggesting its potential as a replacement or addition to the current receptive field mapping block, with Alexander and others agreeing to review the paper and code before the next meeting.
Protocol Introduction and Rationale: Jerome described the Zebra protocol's ability to measure orientation tuning and receptive fields in V1 within five minutes, proposing its evaluation for inclusion in future experiments to save time and improve efficiency.
Next Steps for Evaluation: Alexander and Jerome agreed to review the published paper and available code, inviting team members to provide feedback and discuss its implementation at the next meeting, with plans to coordinate via the discussion forum.