Here we encourage scientists to propose topics and research questions that could be used to build a community proposal. We will aggregate question and ideas at the end of the Question phase into a global poll to assess interest and support of any individual scientific question.
An example of this poll is here : 2023 Community proposal Question Phase : Poll with list of scientific questions
Where should we propose topics and research questions?
1 Like
Right here 
Yesterday, at the OpenScope and Neuropixels technical workshop https://alleninstitute.org/events/neuropixels-and-openscope-workshop/, we had our first brain-storming session for the community proposal. This was for us an opportunity to test the process and get feedbacks on our plan. The community proposal will be more broadly announced to the community within the next week.
We had a one hour long discussion on potential scientific questions and topics that could be at the core of this community proposal. We are extremely thankful to all scientists that attended the workshop and joined the discussion. I hope all can continue the conversation here.
Here is the whiteboard notes we took during the session. I am going to convert my approximate hand-writing into a serie of posts that explain the different ideas we discussed.
In blacks are topics from previous OpenScope projects. In red are new ideas raised during our discussion. Obviously we ran out of space on the whiteboard.
Previous projects :
We first discussed the general topics of previous projects funded by OpenScope:
You can find more information on those on this separate thread: https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
Many of our previous project looked into the neuronal correlates of prediction error and prediction sequences. You can see below that we also explored many other avenues. All those project involved passively viewing mice. To note, we are now accepting behaving project. below I summarize the introduction of these projects that was done during the workshop.
-
Credit assignment project ā This is a project that was funded by the Allen Institute and the Falconwood foundation. The dataset is available as well the current pre-print https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
This project was collected prior to NIH funding. -
Predictive Coding project ā This is a project that was funded by the Allen Institute. The dataset is available https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
This project was collected prior to NIH funding. -
Meaningful project ā Perception / Valence of stimuli across layers and areas. The dataset is available here :
The publication is available https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
This project was collected prior to NIH funding. -
Multiplex project ā This project looked in the role of Excitatory and inhibitory neurons role in errors related to surround suppression. The dataset is currently being converted and will be available this year.
-
Motion project ā This project looked in the encoding of visual motion across the mouse visual hierarchy with two-photon imaging. The dataset is currently being converted and will be available this year.
-
Periodic stimuli (Gamma and others) responses project ā Roles of cell types in encoding stimuli that resonates at specific frequencies. The dataset is available https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
The scientific publication is currently being worked on. -
Illusion project ā Neuronal correlation of optical illusion. This is a project from RFP 2022. The team is establishing illusion as a proxy to study visual perception and binding. The dataset will be available on Dandi. A first pre-print of this work is available https://community.brain-map.org/t/are-there-examples-of-previously-selected-openscope-projects/2412
-
Oddball project ā This is a project from RFP 2022. We are looking into the neuronal correlates of local and global errors across the mouse visual hierarchy using neuropixels recordings. There will be a poster at SFN 2023 on this project.
-
Dendritic coupling project ā This is a project from RFP 2022. We are looking into how dendrites of the layer I are coupled with somas of layer 2/3 and 5 using simultaneous calcium imaging during session with unexpected stimuli.
-
Sequence Learning project ā This is a project from RFP 2023 that is currently ongoing. We are looking into the neuronal correlates of visual sequences across the visual hierarchy, cortical layers using multi-area calcium imaging.
-
Hippocampus and vision project ā This is an RFP 2023 project. We are collecting the data this year. We are looking into visual responses in the cortico-hippocampal network.
-
Bar coding project ā This is an RFP 2023 project. We are collecting the data this year. We are looking at the response of cells across the hierarchy to a very strong salient stimulus that is hypothesized to evoke very stereotypical responses.
1 Like
In the following posts, I will list the potential discussed topics around few thematics.
Feel free to reply and discuss each post.
Brain states projects :
-
Drug effect ā Impact of specify drugs on neuronal responses. This could be done following IP injection.
-
Anesthesia characterization ā A public dataset that provides information on the impact of anesthesia on neuronal activity. This could be compared with all our existing controlled responses.
-
Oxygen levels ā How oxygen levels impact neuronal activity. This could be done by controlling oxygen levels provided through breathing
-
Neuronal correlate of dreaming / state transition / sleeping ā We discussed existing work. A standardized dataset could have broader impact.
1 Like
Predictive and internal models projects:
-
1D navigation dataset ā With mismatch. Similar to (Keller et al, 2012). This would be a data that replicates prior studies results but that is available to the entire community. This involves closed-loop control of the visual stimulus on the screen by the wheel rotation. Some period where the coupling could be turn on and off
-
Predictive coding beyond vision ā Looking the role of higher order / frontal areas.
-
Predictive coding in active mice ā All of our other datasets were with passive mice. We could for instance look into the influence of rewards. Introduce prediction reward errors.
1 Like
Learning projects :
-
Chronic learning recording ā A dataset that provide neuronal responses throughout learning. This is also a large field. The experiment could be designed based on our current task.
-
Habits learning ā This could the simple formation of an automatic motor response. Potentially, a project could look into how such habitual motor response are formed.
Brain-wide projects:
-
Multi-modal learning paired with multi-area recording ā This implies some form of brain-wide recordings.
-
Cross-hemisphere binding ā How neurons across both hemisphere in the mouse work together?
On the cellular diversity in the Brain.
-
Astrocytic imaging ā a dataset that compare astrocytic responses to neuronal responses. It is now established glial cells follow sensory stimuli. We discussed the impact of a shared dataset.
-
Roles of individual cell types in behavioral tasks ā This is a large topic with many existing projects in the community. We discussed this briefly. This is potential a huge topic.
- Major conceptual advance recollection ā We discussed the idea to recollect datasets from major publications that shaped our field so that an ecosystem of datasets could be built. Perhaps this community proposal could be such a dataset.
- Cortical correlate of social communications between mice ā We discussed using stimuli to evoke social responses. Like presenting images of other mice, odors and having other gender mice around.
- optimal stimuli ā using machine learning or other methods to find the stimuli that neurons respond the strongest to.
1 Like
I posted a new topic, but Iām not sure I posted it in the right place. Do you have a way of checking that?
I see the topic created. let me reply there.
Thanks - I also found the OpenScope call for 2023, and see that it does not include testing on basal ganglia neurons, so I suspect that the proposed tests will be outside of the scope of OpenScope. However, I would be glad to discuss the hypothesis and supporting evidence with anyone who might be interested.
Yesterday, we discussed various projects and ideas on how to kickstart the community process.
One idea, suggested by Pamela Reinagel initially, is to start from existing constraints and modify existing task design as a group.
Two option were discussed:
-
Building a shared passively viewing experiment with several sub-blocks presenting stimuli proposed by different groups. For example, we could envision having 10 min blocks of a given stimuli and move on to another stimuli later on. Each block could be designed by another team.
We discussed whether this could create un-wanted impact of stimuli on each other.
It is to be noted that many passively viewing experiment have such basic independence assumptions. -
Start from the detection of change and modify it to tackle specific questions. One team suggested having a passive block intersperse with the active block so as to map many more image presentations.
We discussed modify the task design to bring about different memory components.
One team suggested a design that leverage a lot of natural images to test AI models and their correspondence with different areas across the cortical hierarchy of mice.
The recent ML tool called CEBRA comes to mind on this topic as a powerful mean to test these ideas.
I like the chronic learning idea. I am not aware of a dataset that does that over many brain regions simultaneously in something that is hard enough to learn that you can see learning dynamics occurring over days. I would add to this that doing some form of reversal learning could be a good addition to this project and if possible go beyond learning and try to look at both learning and forgetting in that task. Although that might be a lengthy experimental design.