WormWideWeb-Logo

Encoding summary - CePNEM Model

Explore neuronal encodings of behaviors. Details about the methods can be found here.

Encoding table legend
  • Encoding strength for each of the three behaviors indicates an approximation of the relative variance in neural activity explained by each behavioral variable.
  • Forwardness and Dorsalness capture the steepness of the tuning to velocity and head curvature, respectively.
  • Encoding timescale is the median exponentially weighted moving average (EWMA) half-decay time from the CePNEM model, which indicates how the neuron's activity weighs past versus present behavior for the parameter(s) that it encodes.
  • For all other tunings, the number shows the fraction of neurons that significantly encoded that tuning/feature.

Neuron label legend
  • For neuron class with Dorsal/Ventral (D/V) and/or Left/Right (L/R) differentiation, ? notes that it was not possible to determine that attribute.
  • Example: for RMD class, RMD?L (was not possible to determine D/V). RMD?? (was not possible to determine both L/R and D/V).

Velocity encoding


Head curvature encoding


Feeding encoding


Others
Neuron Count Velocity tuning Head curvature tuning Feeding tuning Timescale (s) Enc. variability
Strength Fwd-ness Fwd Rev Fwd slope + Fwd slope - Rev slope + Rev slope - Fwd slope > Rev slope Fwd slope < Rev slope Strength Dor-ness Dorsal Ventral Dorsal during F Ventral during F Dorsal during R Ventral during R More D during F More V during F Strength Feedingness Act Inh Act during F Inh during F Act during R Inh during R More A during F More I during F


Neuron encoding

Explore neuronal encodings of a given neuron class across different datasets.

Encoding table legend
  • Encoding strength for each of the three behaviors indicates an approximation of the relative variance in neural activity explained by each behavioral variable.
  • Forwardness and Dorsalness capture the steepness of the tuning to velocity and head curvature, respectively.
  • Encoding timescale is the median exponentially weighted moving average (EWMA) half-decay time from the CePNEM model, which indicates how the neuron's activity weighs past versus present behavior for the parameter(s) that it encodes.
  • For all other tunings, the number shows the time segments that significantly encoded that tuning/feature:
    • 1: the neuron encodes the feature in the first time segment
    2: the neuron encodes the feature in the second time segment
    1,2: the neuron encodes the feature in both time segments
    Empty cell: the neuron does not encode the feature

Neuron label legend
  • For neuron class with Dorsal/Ventral (D/V) and/or Left/Right (L/R) differentiation, ? notes that it was not possible to determine that attribute.
  • Example: for RMD class, RMD?L (was not possible to determine D/V). RMD?? (was not possible to determine both L/R and D/V).

Velocity encoding


Head curvature encoding


Feeding encoding


Others
Dataset Neuron Label Velocity tuning Head curvature tuning Feeding tuning Timescale (s) Enc. variability Action
Strength Fwd-ness Fwd Rev Fwd slope + Fwd slope - Rev slope + Rev slope - Fwd slope > Rev slope Fwd slope < Rev slope Strength Dor-ness Dorsal Ventral Dorsal during F Ventral during F Dorsal during R Ventral during R More D during F More V during F Strength Feedingness Act Inh Act during F Inh during F Act during R Inh during R More A during F More I during F