Meeting Notes
- Date: 2025-06-03
- Time: 09:00 (PT)
- Location: Teams Meeting
- Presentations: @Sarruedi @maierav @Dedalus9 @koweiss @jeromelecoq
Agenda
- Brief update to SLAP2 data collection (5 min talk, Jerome)
- Analysis of SLAP2 data by Sarah Ruediger at UCL (10min with discussions)
- Analysis of SLAP2 data by Alex Maier at Vanderbilt (10min with discussions)
- Analysis of SLAP2 data by Nicholas Rodriguez and Karim Oweiss at U Florida (10min with discussions)
- Impact on next experiments on SLAP2 : Discuss potential changes
Meeting Recording
Notes
Meeting Agenda: Jerome outlined the meeting agenda, which includes discussing the analysis of slap 2, summarizing data collection, and presenting analysis from Alex, Sarah, Nicholas, and Lucas.
Agenda Overview: Jerome outlined the meeting agenda, which includes discussing the analysis of slap 2, summarizing data collection, and presenting analysis from Alex, Sarah, Nicholas, and Lucas. He emphasized the importance of keeping time at the end to discuss future experiments.
Presenter Order: Jerome specified the order of presenters: Alex Meyer would go first, followed by Sarah, Nicholas, and Lucas if time permits. Lucas was expected to show statistical analysis pictures.
Data Collection Summary: Jerome summarized the data collection, mentioning that 33 mice were used in three experiments, with imaging input to a single pyramidal cell and recording epical and proximal dendroids during oddball experiments.
Experiment Details: Jerome detailed that 33 mice were used in three experiments, all conducted on the same animal labeled with Glusnan for 4:00. Imaging was focused on a single pyramidal cell, recording epical and proximal dendroids during oddball experiments.
Current Experiment: Jerome mentioned that they are currently conducting an experiment with the same Glusnan for four but now including calcium imaging onto the cells. This experiment was done last week and will likely be repeated this week.
Future Experiments: Jerome emphasized the importance of informing future experiments with the analysis from the current experiments, highlighting the pilot nature of the current experiments to guide future work.
Oddball Experiment Analysis: Alexander presented his analysis of the oddball experiment, explaining the three types of oddballs and their responses, and shared the code and figures in the discussion forum.
Oddball Types: Alexander explained the three types of oddballs: a static grading and two different orientations. These oddballs interrupt the repeated sequence of the same stimulus.
Data Analysis: Alexander discussed the challenge of identifying the oddball experiment parts in the data, as the current data does not specify this. He used code to define the oddball blocks and analyze the responses.
Response Analysis: Alexander found that one oddball type showed no different response compared to the regular stimulus, while another oddball with a different orientation showed a larger response, likely due to orientation tuning. The static stimulus also gave a different response.
Code Availability: Alexander shared the code and figures in the discussion forum, providing a Google collab document with comments and code that can be run in the cloud without installing prerequisites.
Orientation Tuning Analysis: Alexander discussed his analysis of orientation tuning, explaining the response curves and polar plots, and mentioned that the code will be published soon.
Orientation Tuning: Alexander explained that neurons in V1 respond to different orientations with varying response rates. He analyzed two blocks of the experiment where different orientations were shown in random sequence.
Response Curves: Alexander described the response curves as a function of orientation, plotted in polar plots to show tuning sharpness and overlap among synapses.
Code Issues: Alexander mentioned that the plotting function in the code is still buggy, particularly affecting the population average tuning. He plans to release the code soon, even if the bug is not fixed.
Future Analysis: Alexander suggested further analysis on spatial distribution, synchrony, and other aspects, encouraging others to use and expand on the code.
Thresholding and Data Quality: Sarah explained her thresholding criteria for selecting responsive RIs based on C scores, and the need to define criteria for analysis.
Thresholding Criteria: Sarah explained that she thresholded the RIs based on a C score above 0.5 to eliminate those with no evoked response to visual stimuli. This threshold was set arbitrarily and may need refinement.
Data Quality: Sarah highlighted the need to define criteria for selecting RIs for analysis, mentioning that some recordings had a higher C score distribution, but most were around 1.
Response Variability: Sarah emphasized the importance of understanding trial-to-trial variability in responses, rather than just averaging responses to determine significance.
Normalization and Error Estimation: Jerome and Sarah discussed the importance of normalization and error estimation in the analysis, emphasizing the need to aggregate signals before normalization.
Normalization Process: Jerome explained that normalization should be done as the last step of analysis to avoid averaging in noisy data incorrectly. Aggregating signals in the raw photon space before normalization is crucial.
Error Estimation: Sarah and Jerome discussed methods for error estimation, including normalizing at the level of individual stimulus presentations and using propagation of errors.
Data Packaging: Jerome mentioned that the current data packaging is missing some information, such as the F0 values, which are necessary for proper normalization. This will be corrected soon.
Functional Clustering: Sarah presented her approach to clustering synapses based on functional similarity and visualizing their anatomical locations, highlighting the need for refined criteria.
Clustering Approach: Sarah described her method of clustering synapses based on the correlation of their activity patterns, using hierarchical clustering to identify functionally similar groups.
Spatial Visualization: Sarah visualized the anatomical locations of clustered synapses, showing how functionally similar synapses are distributed across the dendritic tree.
Threshold Criteria: Sarah emphasized the need to define refined criteria for clustering, such as setting thresholds for similarity and response strength, to improve the accuracy of the analysis.
Spatial Distribution Analysis: Nicholas shared his analysis of spatial distribution and responsiveness of RIs to different stimulus presentations, and the need to correct for photo bleaching.
Spatial Distribution: Nicholas analyzed the spatial distribution of RIs and their responsiveness to different stimulus presentations, comparing responses across different experimental blocks.
Photo Bleaching: Nicholas highlighted the need to correct for photo bleaching, as it affects the amplitude of responses over time. He used a rolling mean to detrend the signal, but further correction is needed.
Responsiveness Comparison: Nicholas compared the average responsiveness of RIs to stimuli in different blocks, showing how responses change before, during, and after the oddball block.
Similarity Calculation: Nicholas explained his method for calculating similarity between RIs based on orientation tuning vectors and spatial distance, and visualized the results.
Similarity Metric: Nicholas calculated the similarity between RIs using cosine similarity of their orientation tuning vectors, averaged over the 10 nearest neighbors based on spatial distance.
Pre-Post Comparison: Nicholas compared the similarity metrics between the pre and post oddball blocks, showing how the orientation tuning of RIs changes relative to their neighbors.
Next Steps: Jerome proposed continuing the discussion next week, sharing slides on the forum, and focusing on interactions during the meeting. He also mentioned repackaging the data and addressing bleaching correction before making scientific conclusions.
Next Steps and Action Items
Next week’s meeting will focus on script development for the sensory motor oddball