Neuroinformatics�Analysis of multiple spike trains
Kenneth D. Harris, UCL
What drives brain activity?
Sensory stimulus
(controlled by experimenter)
Internal state
Neurons you happen to be recording
Observable behavior
Cross-correlogram
Cortex: putative monosynaptic connections between putative excitatory (red) and FS inhibitory cells (Bartho et al J Neurophys 2004)
CA1: non-synaptic CCGs between putative excitatory cells (Harris et al Nature 2003)
Scaling for cross correlogram
Types of correlation
Sensory stimulus
(controlled by experimenter)
Internal state
Neurons you happen to be recording
Observable behavior
How to compute them
Joint peristimulus time histogram (JPSTH)
https://www.med.upenn.edu/mulab/jpst.html
Conditional independence
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
Cell
Repeat
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
Cell
Repeat
Looks conditionally
independent
Not conditionally
independent
Peer-prediction
Peer prediction for hippocampal place cells
Harris et al Nature 2003
Timescale of peer prediction
Harris et al Nature 2003
Spontaneous activity in mouse V1
Stringer et al Science 2019
https://github.com/MouseLand/rastermap
Shared variance components analysis
Stringer et al Science 2019
What drives brain activity?
Sensory stimulus
(controlled by experimenter)
Internal state
Neurons you happen to be recording
Observable behavior
Multidimensional prediction from behavior video
Stringer et al Science 2019
Stimulus-independent activity is almost orthogonal to stimulus-related activity
Cross-validated PCA to find high-dimensional structure of stimulus-dependent correlations
Stringer et al Nature 2019
Power-law eigenspectra
Stringer et al Nature 2019
Power-law eigenspectra
Stringer et al Nature 2019