It’s in there, check out the latest data loader.
2 Likes
thank you a lot, i didn’t notice it
1 Like
That would be INSANELY helpful.
Hi!
I was wondering if the neural spike data is from all the neurons recorded during a given session or does it consist on the neurons that are significantly modulated during the task?
Thank you!
All the neurons that fired at least one spike during the task.
1 Like
If I label the spike bins from 0ms to 2500ms, is it correct to say that the stimulus onset is at 500ms and the ‘gocue’, ‘feedback_time’, and ‘response_time’ values can be read as is from the data? In other words, is this the proper zero time point to use to accurately place all the event times given in the data?
5 Likes
Hello!
I want to access to the LFP data in python. Has anybody already done this?
I downloaded the raw data and after compressing the file I have an *.imed.lf.bin that it turns out to be a Neuropixel.ImecDataset, and I don’t have any clue how to handle it.
+1 for this and also the script. I think raw spike_times and activity during ITI (or just the interval after feedback to the start of next trial) would be extremely helpful. We can do the spike counts ourselves.
1 Like
Hello! My group is exploring how pre-stimulus activity affects judgement accuracy and we were figuring out how to navigate the colab notebook. Could anyone explain what are the 5 grey plots under the cell named ‘correct vs incorrect trials’? How exactly are the trials sorted by latency (which my group is assuming to be referring to response time)? Thank you!
Hello, my group is looking at the correlation between cells of different regions. We were wondering how do we get the brain location for a given spike cluster (from the actual dataset, and not the pre-processed dataset provided). There is this ‘channels.brainLocation.tsv’ provided in the dataset, however we noticed that the number of entries is different from the number of unique spike clusters. Hence, if we are not wrong, we need additional information to relate the spike clusters to this brain location dataframe. Does anyone have any idea on how to approach it?
Thanks!
Is there a list of all the brain region identifiers? like the names for the regions with the code of the brain area?
No, everything is in relation to stimulus onset, which itself happens at 500ms on every trial.
2 Likes
The LFP data is already included in the curated data. Take a look at the latest data loader to see how to use it.
1 Like
Notice in that cell it says the trials are sorted by “correct” first, and then by latency. This is achieved by sorting the array is_correct + response/10. In this case, response is the detected movement after go cue onset, but usually it’s not the first movement the mouse makes.
There is this ‘channels.brainLocation.tsv’ that provides all the brain region identifiers, i.e. allen_ontology. But i’m not too sure how to relate it to spike clusters
What kind of codes are you looking for? If you look through the notebook, there is a cell that defines all the regions and brain_groups. This information can be matched with dat[‘brain_area’] for each neuron.
Yes, we were also wondering what ‘root’ refers to: it’s in the brain_are key of the dat dictionary.
Thanks in advance for any help 
2 Likes
Ah, I think those are just channels outside of the brain, or on the edge of the brain. If there are neurons there (there shouldn’t be very many), they are likely noise or there’s something weird with them, or they’re just classified as “other” in Nick’s classification.
3 Likes
- How do we convert AP-DV-LR coordinates given in channels.brainLocation.tsv to world coordinates/other coordinates. We are looking to cluster neurons even inside brain_lfp_areas. I could find AP-ML-DV(http://labs.gaidi.ca/mouse-brain-atlas/) coordinate system but not for this
- How do we relate clusters.probes to channels.probe. They have unequal number of [0, 1, 2] probes.
cluster.probes (array([0., 1., 2.]), array([421, 339, 459]))
channels.probes (array([0., 1., 2.]), array([374, 374, 374]))
EDIT: 2nd question is about 11th experiment
1 Like
Here is what Nick says: You just need the ccf atlas. It contains the “annotation volume” which is a 3D matrix of the CCF space where each voxel has an integer corresponding to the brain region at that location. Then you need the “structure tree” which is a text table that tells you the metadata about all those integers, like their name and parent regions and such. They can read about this here: https://github.com/cortex-lab/allenCCF
For channel numbers, the cluster numbers are assuming the probes are stacked on top of each other, so you’d have to take %374 to get channel number relative to each probe. I can post my dataset curation functions and you can dig into those if you want. Alternatively, follow the information on Nick’s wiki as closely as possible.
3 Likes