Somatodendritic consistency check for temporal feature segmentation (Asabuki & Fukai 2020)

 Download zip file 
Help downloading and running models
Accession:263246
"The brain identifies potentially salient features within continuous information streams to process hierarchical temporal events. This requires the compression of information streams, for which effective computational principles are yet to be explored. Backpropagating action potentials can induce synaptic plasticity in the dendrites of cortical pyramidal neurons. By analogy with this effect, we model a self-supervising process that increases the similarity between dendritic and somatic activities where the somatic activity is normalized by a running average. We further show that a family of networks composed of the two-compartment neurons performs a surprisingly wide variety of complex unsupervised learning tasks, including chunking of temporal sequences and the source separation of mixed correlated signals. ..."
Reference:
1 . Asabuki T, Fukai T (2020) Somatodendritic consistency check for temporal feature segmentation. Nat Commun 11:1554 [PubMed]
Model Information (Click on a link to find other models with that property)
Model Type: Dendrite;
Brain Region(s)/Organism:
Cell Type(s):
Channel(s):
Gap Junctions:
Receptor(s):
Gene(s):
Transmitter(s):
Simulation Environment: Python;
Model Concept(s): Learning; Simplified Models; Synaptic Plasticity;
Implementer(s):
/
MRIL_codes
README.txt
bss.py
chunking.py
patternMulti.py
patternSingle.py
                            
These codes are written in python3. We ran the codes with numpy 1.15.3, scipy  1.1.0 and matplotlib 3.0.1.
 For each code, the simulation will finish within 30 minutes on a normal desktop computer.

・patternSingle.py : Detecting spike pattern with single dendritic neuron. Running this code generates figures for input raster plots and the trained output. Example results are shown in Fig. 1B,C.

・patternMulti.py : Detecting multiple spike patterns with dendritic neurons. Running this code generates figures for input raster plots, all trained output activities, trained lateral inhibitory connections and example traces of outputs. Example results are shown in Fig. 2D,E.

・chunking.py : Learning chunks by dendritic neurons. Running this code generates figures for all trained output activities, trained lateral inhibitory connections and example traces of outputs. Example results are shown in Fig. 4C.

・bss.py : Performing blind source separation with simple example. Running this code generates figures for true and mixed sources and trained activities. This is a code for Fig. 7, yet here we used simple input signals since we used external public data sets in Fig. 7. 


Loading data, please wait...