Skip to content

Neural Ensembles 2022#

Assorted notes and thoughts#

Zebrafish imaging with Mishra Ahrens#

  • Whole brain cellular light sheet imaging in zebrafish (120000 neurons, 20000 glia) - see Vladimirov Nature Methods 2014
  • They performed online data analysis on the cluster.
  • Zebrafish swim against the waterflow, but remember where they are.
  • By ablation of cells in SLO-MO, the memory is gone.
  • However, activation of cells in SLO-MO mimics memory and makes the fish swim.
  • SLO-MO projects to the inferior olive (GABAergic - inhibitory cells)
  • Yang et al. 2021 Biorxiv
  • SLO-MO might have an analogy in mammals.
  • About 300 cells in area SLO-MO, which are part of a multi-region network. Ablating Inferior olive - SLO-MO still functions.
  • SLO-MO -> IO -> Cerrebellum -> Motor area -> (the encoded signal becomes closer and closer to behavioural output).
  • IDEA: Ablate random number of cells until you lose the memory (what would be % be and what would the robustness be).

Control of neural networks with high precision - Valentina Emiliani#

  • Classically optogenetic activation turns on many cells at the same time.
  • Trying to transition from whole region optogenetics to circuit optogenetics to manipulate single targets (or a few targets of choice).
  • Holographic illumination (light shaping).

Geometry of neural activity - Carsen stringer#

  • Ahrens at al 2013 (zebrafih)
  • RNA sequencing (Tasic at al 2018, see kobak and berens 2019)
  • FLYEM team
  • Behavioural data (lots of it)
  • Variance vs PC dimensions (eigenspectrum) - it follows a power-law decay, top few PC dimensions are not enough.
  • t-SNE or autoencoders for unsupervised approaches. or the clustering output.
  • The resulting method from this is Rastermap.
  • Step 1 - cluster the data with k-means
  • Step 2 - Rearrange clusters by similarity - sort the matrix of similarities between cluster activity. (keep power-law decay of PC variances) - iterative optimisation (NP HARD)
  • Step 3 - Upsample cluster centers (lin interp I think)
  • Step 4 - Assign neurons and sort (done)
  • Benchmarks with neighbour metrics (local score) and also intermediate and global score of distances. [the idea is neighbours in Nspace vs those in linspace or 3dspace]
  • Can visualise the behavioural predictions of (behaviour estimate neural activity).
  • Ahrens lab zebrafish data - can colour neurons in the brain.
  • Churchland lab - visualise widefield imaging data.
  • Can compute clusters in lagged time
  • Rastermap taking a space filling curve through the data?
  • Check out the rastermap GUI (how does it load the data?)
  • For CI imaging they cluster the deconvolved data. (long decay time of the sensor, you can get rid of the decay by deconvolution - increase SNR due to bleedthrough)
  • It is integrated into Suite2p.

Reading the mind of a Cnidarian - Rafeal Yuste#

  • CNidarians (no central brain) - bilaterians (a bilateral brain)
  • Genetic code was broken in (codons)
  • Hydra vulgaris (200 - 2000 neurons) with 11 neuronal cell types, two nets formed.
  • Can see evry spike from every neuron in such a behaving animal with calcium imaging.
  • Spontaneous correlated (ryhtmic) activity during relaxing.
  • It seems there are about 12 ensembles - they are non-overlapping.
  • Dupre and Yuste 2017.
  • Can also image the activity of every muscle cell in the ectoderm and endoderm.
  • gcamp in one, rcamp in the other circuit.
  • Neural activity, muscle activity, behaviour (all can be measured).
  • So one can link all the matrices (adrienne fairhall currently working on this).
  • Goal is to explain how all the 12 ensembles work.
  • You can also take apart the animal and it will reform. (it self assembles)
  • Contraction bursts.

David Anderson#