TY  - JOUR
AU  - Wybo, Willem A. M.
AU  - Tsai, Matthias C.
AU  - Tran, Viet Anh Khoa
AU  - Illing, Bernd
AU  - Jordan, Jakob
AU  - Morrison, Abigail
AU  - Senn, Walter
TI  - NMDA-driven dendritic modulation enables multitask representation learning in hierarchical sensory processing pathways
JO  - Proceedings of the National Academy of Sciences of the United States of America
VL  - 120
IS  - 32
SN  - 0027-8424
CY  - Washington, DC
PB  - National Acad. of Sciences
M1  - FZJ-2023-03213
SP  - e2300558120
PY  - 2023
AB  - While sensory representations in the brain depend on context, it remains unclearhow such modulations are implemented at the biophysical level, and how processinglayers further in the hierarchy can extract useful features for each possible contex-tual state. Here, we demonstrate that dendritic N-Methyl-D-Aspartate spikes can,within physiological constraints, implement contextual modulation of feedforwardprocessing. Such neuron-specific modulations exploit prior knowledge, encoded instable feedforward weights, to achieve transfer learning across contexts. In a network ofbiophysically realistic neuron models with context-independent feedforward weights,we show that modulatory inputs to dendritic branches can solve linearly nonseparablelearning problems with a Hebbian, error-modulated learning rule. We also demonstratethat local prediction of whether representations originate either from different inputs,or from different contextual modulations of the same input, results in representationlearning of hierarchical feedforward weights across processing layers that accommodatea multitude of contexts.
LB  - PUB:(DE-HGF)16
C6  - 37523562
UR  - <Go to ISI:>//WOS:001121663700001
DO  - DOI:10.1073/pnas.2300558120
UR  - https://juser.fz-juelich.de/record/1014283
ER  -