Meeting Notes

• Date: 2026-03-03
• Time: 09:00AM (PT)
• Location: Teams Meeting

Agenda

Discussion on how to organize and focus future analysis:

• Discussion of analysis poll to be sent to all. Draft is here
• Analysis plan graphic introduced by @Sarruedi

Meeting Recording

Meeting Notes

Poll Structure for Analysis Coordination: Jerome outlined the proposed structure of a poll to coordinate analysis efforts, aiming to gather information on participants' expertise, prioritize analysis threads, and group analyses for publication, with input from Lucas, Alexander, David, and others.

Poll Content and Purpose: Jerome described the poll as being based on the aims and analysis questions from their archive paper, structured into three main sections: participants' expertise and preferred contributions, ranking analysis sections by scientific impact and personal interest, and grouping analyses into publication bundles as suggested by Karim. The poll also includes a free-form comment section for additional feedback.

Visual Analysis Map: Lucas asked about the role of a visual analysis map, and Jerome clarified that Sarah had created a complementary visual intended to help organize and show dependencies between analyses, which could be iterated upon before sending the poll.

Iterative Development of Poll and Visuals: Participants discussed whether to finalize the main analysis areas before gathering input or to allow for dynamic grouping, with consensus leaning toward iterating both the poll and the visual map to reflect ongoing contributions and organizational needs.

Analysis Workflow Organization and Visualization: Lucas, Alexander, Stefan, and Jerome discussed strategies for organizing the analysis workflow, including the use of visual tools, breaking down tasks into preprocessing and analysis layers, and ensuring transparency and coordination to avoid duplicated efforts.

Top-Down vs. Bottom-Up Approaches: Alexander highlighted differing philosophies in scientific projects—some preferring a top-down, goal-oriented approach, others a bottom-up, stepwise method—and suggested tools like Kanban boards or Gantt charts to clarify task dependencies and expectations, especially for trainees.

Visual Representation of Workflow: Lucas proposed using a shareable Google Slides document to visually map out analysis projects, with central nodes for existing code and preprocessing pipelines, and links to more specialized analyses, allowing contributors to see both the big picture and available resources.

Layered Analysis Structure: Stefan and others agreed on a three-level structure: level 0 for preprocessing and quality control, level 1 for basic characterization, and level 2 for analyses addressing main scientific questions, with flags rather than deletions for data selection to maintain flexibility.

Coordination and Transparency: David raised concerns about the number of active contributors and potential duplication of effort, prompting Jerome to emphasize the importance of clear action items, transparency, and a flexible, open process to facilitate participation and coordination.

Authorship, Credit, and Contribution Tracking: Nicholas A Rodriguez, Alexander, Stefan, and Jerome discussed concerns about authorship attribution, the use of tools like GitHub for tracking code contributions, and the need for transparent contribution tables in publications.

Concerns About Lost Credit: Nicholas A Rodriguez expressed concerns that original contributions to code and analysis might be lost as projects evolve, suggesting the use of GitHub's authorship tracking features to ensure proper credit.

Transparent Contribution Tables: Jerome proposed that each publication include a clear contribution table specifying individual roles and code contributions, with author order to be jointly agreed upon by all coauthors, ensuring transparency and recognition.

Challenges in Large Collaborations: Alexander noted the practical difficulties in tracking contributions over time, especially as participants move on, and emphasized the importance of ongoing engagement and clear documentation to maintain accurate authorship records.

Evolving Authorship Norms: Alexander and Stefan discussed the changing landscape of scientific credit, with increasing emphasis on metrics like citations and H-index, and the potential for large, multi-author papers to provide significant career benefits.

Data Publication and Analysis Scope: Jerome, Lucas, Stefan, and Alexander planned the creation of a data release publication encompassing all modalities, discussed the scope of analyses to include, and agreed on the importance of standardizing basic analyses across datasets.

Data Release Publication Plan: Jerome outlined the plan to publish a data release paper after data collection is complete, with DANDI set DOIs for each modality and NWB files containing preprocessed data, inviting all contributors to be coauthors.

Scope of Included Analyses: Lucas and Stefan discussed including level 0 and 1 analyses (preprocessing and basic characterization) in the data paper, while reserving more advanced, question-driven analyses for subsequent publications.

Standardization Across Modalities: Alexander suggested that basic analyses, such as identifying visually responsive signals and orientation tuning, should be performed for all modalities to ensure consistency and provide a foundation for further research.

Potential for Multiple Publications: The group considered whether to publish a single comprehensive data paper covering all modalities or separate papers for each, with consensus leaning toward a unified publication for greater impact.

Next Steps and Action Items: Lucas agreed to draft a Google Slides visualization of the analysis workflow, with plans to iterate on it in the following week, and Jerome emphasized the need to involve Sarah and maintain open communication as the project progresses.

Drafting Workflow Visualization: Lucas committed to creating a draft Google Slides document based on the existing PDF and prior experience, mapping out code contributions and analysis steps for each dataset, to be reviewed and refined by the group.

Involving Key Contributors: Jerome highlighted the importance of including Sarah, who created the initial analysis graphic, in ongoing discussions and iterations of the workflow visualization.

Scheduling Follow-Up: The group agreed to revisit and iterate on the workflow visualization in the next meeting, allowing time for careful planning as data collection continues.