# Open Data Symposium

## Outline
Date: Oct
Timeframe: 2.5 days
Public: Inivitees (PIs, Senior Staff), open to students/postdocs
Attendees: aim for 50+ positive RSVPs
Location: MIT McGovern Brain Research Institute
43 Vassar St, Cambridge, MA 02139, USA
Area coverage: US-wide

## Topics:

### I. Data Archives
Data archives to provide a short summary of their rationale and capabilities — if time allows, highlighting great work done with their data.
* Dedicated infrastructure:
* BossDB
* NeMO
* BIL
* DANDI
* OpenNeuro-reliant
* OpenNeuro
* NEMAR
* OpenNeuro-PET
* DABI

### II. Data Sources
* Measurement Devices (open data acquisition question — e.g. openephys OR Bruker+NIfTI)
* High throughput acquisition (and resulting big data challenges — e.g. Ca Imaging/Janelia OR Connectome/Harvard, MultielectrodeArrays/Maxwell Biosystem, Neuropixels)

### III. Research Output
* Software (e.g. spikeinterfaces; fmriprep; pycortex-galant; MNE; EBRAINS-suite)
* Analysis (e.g. Dimensionality Reduction AND/OR Multiscale integration)
* Discovery Workflows (how do open data elements integrate in the overarching research process — e.g. Scientific Machine Learning ecosystem AND/OR Physics-informed NN for data analyses)

### IV. Big Picture & Challenges
* Search
* Open data ethics (Privacy, Resilience, and Sustainability — e.g. FSF OR OSI)

## Schedule
- Panel discussion for each session on challenges.
- Only limited Q&A after individual talks
- Full discussion in the panel discussion.

### Day 1; Archives examples
9:00AM-5.30PM
- 09:00–9:15 Coffee/Tea
- 9:15–9:30 Wellcome to symposium
- Session I.a
- 9:35–9:40 Intro
- 9:40–10:00 BossDB example
- 10:00–10:20 NeMO example
- 10:20–10:40 BIL example
- 10:40–11:00 DANDI example
- 11:00–11:30 Panel discussion
- 11:30–13:00 Lunch Break
- Session I.b
- 14:05–14:10 Intro
- 14:10–14:30 OpenNeuro example
- 14:30–14:50 NEMAR example
- 14:50–15:10 OpenNeuro-PET example
- 15:10–15:30 DABI example
- 15:30–16:00 Panel discussion
- 16:00–16:05 Outro, thanks
- 16:15–17:30 Discussion on ethics/concerns of data sharing
- 18:00–20:00 Dinner and Drinks with the Guest Speakers

### Day 2; Data/Measurements
9:00AM-5.30PM
- 09:00–9:30 Coffee/Tea
- Session II.Measurement Devices (OpenEphys,EM/connectiome,MesoSpim,Neuropixel)
- 9:35–9:40 Intro
- 9:40–10:00 Talk 1
- 10:00–10:20 Talk 2
- 10:20–10:40 Talk 3
- 10:40–11:00 Talk 4
- 11:00–11:30 Panel discussion
- 11:30–13:00 Lunch Break
- Session III. High throughput acquisition (Voltage imaging, spatial transcriptomics, MEA, FIL)
- 14:05–14:10 Intro
- 14:10–14:30 Talk 1
- 14:30–14:50 Talk 2
- 14:50–15:10 Talk 3
- 15:10–15:30 Talk 4
- 15:30–16:00 Panel discussion
- 16:00–16:05 Outro, thanks
- 16:15–17:30 Discussion with archive representatives on challenges and solutions
- 18:00–20:00 Dinner and Drinks with the Guest Speakers

### Day 3; Methods/Analyses
9:00AM-5.30PM
- 09:00–9:30 Coffee/Tea
- Session II.(Behavior automation,Neurophys tools, Human Phys)
- 9:35–9:40 Intro
- 9:40–10:00 Talk 1
- 10:00–10:20 Talk 2
- 10:20–10:40 Talk 3
- 10:40–11:00 Talk 4
- 11:00–11:30 Panel discussion
- 11:30–13:00 Lunch Break
- Session III. (Cell, Circuit, Gene)
- 14:05–14:10 Intro
- 14:10–14:30 Talk 1
- 14:30–14:50 Talk 2
- 14:50–15:10 Talk 3
- 15:10–15:30 Talk 4
- 15:30–16:00 Panel discussion
- 16:00–16:05 Outro, thanks
- 16:15–17:30 Discussion on search
- 18:00–20:00 Dinner and Drinks with the Guest Speakers

### Day 4; Future/Challenges
9:00AM-5.30PM
- 09:00–9:30 Coffee/Tea
- Session II.(New Frontiers, High Dimensional data, Multiscale integration)
- 9:35–9:40 Intro
- 9:40–10:00 Talk 1
- 10:00–10:20 Talk 2
- 10:20–10:40 Talk 3
- 10:40–11:00 Talk 4
- 11:00–11:30 Panel discussion
- 11:30–13:00 Lunch Break
- Session III. (Challenges: HPC, Reproducibility, Scientific Data Standards, Scaling)
- 14:05–14:10 Intro
- 14:10–14:30 Talk 1
- 14:30–14:50 Talk 2
- 14:50–15:10 Talk 3
- 15:10–15:30 Talk 4
- 15:30–16:00 Panel discussion
- 16:00–16:05 Outro, thanks
- 16:15–17:30 Discussion on Resilience and Sustainability
- 18:00–20:00 Dinner and Drinks with the Guest Speakers

## Potential speakers

Measurement Devices
- Bruker
- inquire for speaker (Horea)
- Fabian Voigt (Harvard/MesoSPIM → http://mesospim.org/)
- Josh Siegel (Allen): open ephys
- Jacob Vogt (Janelia): open ephys
- Jeff Litchman (Harvard): connectome/brainbow
- Davi Bock (Vermont): Connectome (BRAIN)
- Ken Harris (UCL)/Svoboda (Allen): Neuropixel)

High throughput acquisition
- Ed Boyden (MIT): Spatial multiplexing flourescent reporters
- Adam Cohen (Harvard): Voltage Imaging
- Marius Pichitaru (Janelia): high-throughput acqusition
- Urs Frey (Maxwell Biosystems): MEA high-throuput acquistion
- Andreas Hierlemann (ETHZ): MEA high-throuput acquistion
- Ellizabeth Hillman (Columbia): Optical imaging
- Alipasha Vaziri (Rockefeller): Bead Imaging
- Bernardo Sabaitini (Harvard): Super-resolution 2-photon + Fluorescence-lifetime imaging microscopy

Behavior; automated analyses
- Mackenzie Mathis (EPFL): Cebra; Learnable latent embeddings for joint behavioural and neural analysis
- Bob Datta (Harvard): MoSeq, Behavioral quantification
- Kristin Branson (Janelia): machine vision; insect behavior tracking

Neurophysiology tools and analyses
- Adrien Peyrache (McGill): Pyneapple/NWB, Spatial coding, Head direction cell, manifolds
- Loren Frank (UCSF): Hippocampal replay, point process/mixture models, DANDI
- Scott Lindermann (Stanford): whole brain imaging/Single trial stat
- Jeremy Magland (Flatiron CCM): high throughput spike sorting; Ironclust/mountainsort/Spiking Interface

High dimensional data, neurophysiology/neuroimaging
- Alex Williams (Flatiron CCN): Unsupervised discovery of low-dimensional dynamics
- Bing Brunton (UW): Empirical dynamic mode decomposition, iEEG, DANDI
- Ben Fulcher (Sydney): ComputeEngine, Highly comparative time-series analysis (fMRI/MEG)
- Jack Galant (Berkeley): pycortex, ML/Feature space selection (fMRI)
- Ole Jensen (Birmingham): Oscilllation, Flux pipeline (OpenNeuro,MEG)

Disease [Alzherimer's, Parkinson, ...]
- NEMAR [to add]
- OpenNeuroPET [to add]
- BIL [to add]

ML, new frontiers (Scientific workflow)
- Chris Rackakaus (MIT): scientific machine learning
- Karen Willcox (UT Austin): Dimensionality reduction, data-driven reduced models
- Petros Koumoutsakos (Harvard): uncertainty, optimization for multiscale models

OpenScope (DANDI)
- Joel Zylberberg (York): Allen OpenScope, Soma/dendrite computation
- Luca Mazzucato (Oregon): Allen Institute Openscope - Differential encoding of temporal context and expectation
- Anton Arkhipov (Allen): Allen Institute Openscope - Measuring Stimulus-Evoked Neurophysiological Differentiation in Populations of Neurons

Neural circuit and computation (BossDB)
- Marta Zlatic/Albert Cardona (Janelia): BossDB, Dorsophila circuit
- Reiner Fredrich (FMI): BossDB, Zebrafish olfactory
- Andreas Tolias (Baylor/Stanford): BossDB/DAND, MICrONS, Interneuron

Human invasive physiology (DABI)
- Fedrenko (MIT): language processing, iEEG
- Ueli Rutishauser (Cedar Sinai): single unit, memory
- Ed Chang (UCSF): speech motor cortex

Gene/Cell (BIL/NEMO)
- Guoping Feng (MIT): spatial transcriptomics; BIL
- Hongkui Zeng (Allen): cell morphology; BIL
- Bosiljka Tasic (Allen): cell type; BIL
- Hong-Wei Dong (UCLA): cellular connectivity; BIL
- Yongsoo Kim (Penn state): receptor circuit mapping
[@Satra to add NEMO]

Archive representatives
- BossDB
- Erik Johnson, APL
- NeMO
- Owen White, UMD
- BIL
- Alexander Ropelewski, Pittsburgh Supercomputing Center
- DANDI
- Yarik
- OpenNeuro:
- Russell Poldrack, Stanford
- NEMAR (Uses openneuro):‪
- Scott Makeig, UCSD
- OpenNeuro-PET (uses openneuro):
- Robert Innis, NIH
- DABI:
- Dominique Duncan, USC LONI

## Invitations

### Pre-invitees List

Rationale: Have a number of confirmed speakers for when we send invitations out to people who may decline out of prominence concerns.

### Invitees List

Send out an email with a link to the schedule (sessions and confirmed speakers)

## Important considerations

### Funding

Available for speakers and organizers, **not** available for attendees.

### Location: Auditorium at MIT TBD (Nima)
- singleton auditorium for the symposium talks

### Later

a. Communications platform (mattermost, slack, element?)
b. Landing Page — Nima can make the domain, Christian can make the page (opendatasymposium.mit.edu).

# Progress

0. TODO Determine exact date (ST, ND)
1. TODO Write example invitation email, substantiating the focus (HC)
2. TODO Make logo
3. TODO determine room availability and annotate rooms and location above (ND,SG,DJ)