Meeting Notes

• Date: 2026-01-13
• Time: 09:00AM (PT)
• Location: Teams Meeting

Agenda

1. follow-up on Receptive Field characterization in pilot: Gratings vs Zebra by @lrudelt and @RaabeM

Meeting Recording

Meeting Notes

Receptive Field Mapping Analysis and Troubleshooting: Lucas and Marcel presented their analysis comparing receptive field mapping using local gratings and a wavelet-based approach, with input and troubleshooting suggestions from Sophie, Max, Alexander, Stefan, and Jerome, focusing on unexpectedly low correlation values and possible causes such as analysis pipeline issues, timing alignment, and stimulus differences.

Comparison of Mapping Methods: Lucas explained the comparison between receptive field mapping using local gratings and a wavelet-based analysis developed by Sophie and Max, highlighting that the expected correspondence between the two methods was not observed, as the wavelet-based correlations were much lower than anticipated.

Potential Causes for Low Correlation: Sophie and Lucas discussed possible reasons for the low correlation values, including the influence of surround effects in the zebra noise stimulus, issues with the computation of correlations, and the possibility of problems in the analysis pipeline rather than the physiological data itself.

Timing Alignment and Data Integrity Checks: Marcel and Jerome described the methods used to align stimulus frames with neural data, including logging frame numbers and timestamps, and discussed the need to verify that the saved stimulus matches what was actually displayed, with Alexander suggesting sanity checks such as reconstructing the movie from saved frames.

Suggestions for Debugging and Additional Experiments: Sophie and Max recommended running experiments with repeated zebra noise presentations to assess repeatability and alignment, and suggested varying wavelet parameters and response windows in the analysis to improve correlation results; they also proposed using spike-triggered averages and comparing responses across repeats for further debugging.

Action Items and Next Steps: The group agreed to spend the coming days reviewing the alignment process, consider modifying the stimulus protocol to include repeated zebra presentations, and continue analysis and discussion offline, with Jerome planning to document the proposed changes and solicit further feedback on the forum.

Experimental Design Decisions for Upcoming Data Collection: Jerome, Marcel, Lucas, Alexander, Sophie, and Max discussed adjustments to the experimental protocol for the upcoming production phase, debating the inclusion and duration of zebra and trippy stimuli, the importance of repeat presentations, and the need to ensure data integrity before proceeding.

Stimulus Protocol Adjustments: The team considered replacing or repeating the zebra and trippy stimuli to improve repeatability and analysis reliability, with suggestions to double the zebra presentation or split sessions between zebra and trippy, ultimately favoring repeated zebra presentations to strengthen statistical power and facilitate debugging.

Balancing Session Duration and Data Quality: Sophie and Jerome discussed the trade-off between session length and the reliability of neuronal responses, agreeing that longer or repeated presentations are preferable for capturing responses from sparsely firing neurons, but acknowledging practical constraints on experiment duration.

Ensuring Data Integrity Before Production: Jerome emphasized the priority of verifying timing and alignment for the zebra stimulus before starting large-scale data collection, stating that any issues with data collection are difficult to correct post hoc, and that the pilot phase should ensure all protocols are robust.

Consensus on Stimulus Inclusion: The group reached consensus to retain the Gabor stimulus as a control and to keep the zebra stimulus in the protocol, provided that alignment and timing are confirmed, with Jerome planning to document this decision and invite further input from the team.

Technical Explanations and Methodological Clarifications: Throughout the meeting, Lucas, Marcel, Sophie, and Jerome provided detailed explanations of the technical methods used for stimulus presentation, data alignment, and analysis, addressing questions from Alexander and Stefan regarding the specifics of the pipeline and the applicability of the wavelet method to different data types.

Stimulus Presentation and Logging: Jerome described the use of Bonsai to load and play movie stimuli, logging both movie frame numbers and session frame numbers, and tracking frame display times at high temporal resolution to ensure accurate alignment with neural recordings.

Applicability of Wavelet Analysis: Stefan inquired about the use of the wavelet-based analysis on electrophysiological data, with Sophie confirming that while there is no preprint, the method has been tested on neuropixel data and works well, though most published results are from calcium imaging.

Analysis Parameter Choices: Sophie and Marcel discussed the selection of wavelet parameters, such as spatial frequency and orientation, and the impact of restricting these parameters to match the Gabor stimulus, noting that relaxing these constraints may improve analysis outcomes.