Editorial: Sensory Adaptation

Mehdi Adibi, Davide Zoccolan, Colin W. G. CliffordFrontiers in Systems Neuroscience 2021 · DOI · OpenAlex

Abstract

How does sensory adaptation relate to the intrinsic properties of neurons versus their connectivity within a neuronal network? This question underpins the theme of two research articles in this research topic. Using in vitro whole-cell electrophysiology from neurons in the auditory midbrain of embryonic birds, Malinowski et al. identified a link between the rapid adaptation behaviour of neurons and their intrinsic physiological properties but not their morphological classification. Phasically spiking neurons exhibiting strong adaptation with faster recovery were found to have higher thresholds, lower membrane resistance and lower membrane time-constants compared to tonically spiking neurons. These two populations exhibited few dendritic morphological differences. Conversely, neurons with different dendritic morphological structures (stellate versus elongated neurons) showed no significant difference in their intrinsic properties and spiking patterns. Quiroga et al., taking a computational approach, predicted that recurrent connections in a population of orientation selective neurons can give rise, over short time scales of a few hundred milliseconds, to attractive perceptual aftereffects instead of commonly reported repulsive aftereffects. They further employed a psychophysical paradigm, appositely designed to capture aftereffects over short time scales, in order to verify this prediction. Their results indicate how recurrent network dynamics can contribute to shape perceptual adaptation.

Introduction

Organisms inhabit environments characterized by extreme fluctuations in sensory input. A visual scene may vary in luminance by several orders of magnitude, and social interactions require the precise tracking of subtle facial cues. To process this information efficiently without saturating signaling capacity, sensory systems must constantly recalibrate. Sensory adaptation serves as the brain's universal "auto-gain" feature, allowing neurons to adjust their sensitivity based on recent history. This editorial synthesizes a multidisciplinary collection of research exploring whether adaptation arises from the internal "hardware" of individual neurons (intrinsic properties) or the way they are "wired" together (network connectivity).

Mechanisms of Adaptation: Intrinsic Physics vs. Morphology

A central question in neurophysiology is whether a neuron's function is dictated by its shape or its internal physics. In the auditory midbrain of embryonic birds, Malinowski et al. demonstrated that a neuron's ability to adapt is tied strictly to its intrinsic physiological properties rather than its morphological classification. Specifically, "phasic" neurons—which exhibit strong adaptation and rapid recovery—possess higher thresholds, lower membrane resistance, and lower membrane time-constants compared to "tonic" neurons. Interestingly, these two populations showed few dendritic morphological differences. Conversely, neurons with distinct structures (stellate vs. elongated) showed no significant differences in spiking patterns, suggesting that intrinsic physiological tuning, rather than physical structure, is the primary driver of adaptation in this system.

Network Dynamics and Perceptual Aftereffects

While adaptation is often associated with "repulsive" aftereffects—such as the motion aftereffect where stationary objects appear to move in the opposite direction of a previously viewed stimulus—computational modeling by Quiroga et al. suggests a more complex reality. Their work indicates that recurrent network connections among orientation-selective neurons can generate "attractive" perceptual aftereffects over very short timescales (a few hundred milliseconds). By combining modeling with a specialized psychophysical paradigm, they verified that recurrent dynamics play a critical role in shaping how we perceive changes in the environment over time.

Hierarchical Specificity and Predictive Coding

Adaptation occurs across various levels of the sensory hierarchy, from elementary motion detectors to high-level social processing:

  • Tactile Systems: In the rodent barrel cortex, adaptation in upper layers is highly specific to individual whiskers. This specificity is not merely inherited from input layers but emerges from horizontal connections within the cortex, facilitating high responsiveness in complex tactile environments.
  • Visual Motion and Faces: Research in blowflies revealed that global motion adaptation is direction-dependent, while local adaptation is not, allowing for robust local movement detection regardless of the animal's heading. In humans, adaptation extends to complex social cues, such as the variance of facial configurations, biasing the perception of emotion and eye gaze.
  • Attention and Hierarchy: Tonelli et al. found that size adaptation in humans is independent of the amount of attention paid to the stimulus. This suggests that size adaptation occurs early in the visual hierarchy, likely in the primary visual cortex (V1).

A recurring functional theme is predictive coding. By suppressing responses to repetitive, predictable stimuli, adaptation reduces "prediction error." This effectively highlights deviant or novel stimuli, acting as a built-in filter for environmental salience.

Conclusion and Future Outlook

The research presented here highlights that sensory adaptation is a sophisticated process driven by both intrinsic cellular physics and complex recurrent dynamics. However, gaps remain in our understanding of how neuromodulators (e.g., dopamine or serotonin) tune these adaptation levels based on internal states. Furthermore, the transition from short-term adaptation (milliseconds) to long-term plasticity (learning) remains a fertile ground for future investigation to determine where the boundary between these two fundamental processes lies.

Summary updated

Abstract

to their connectivity within a neuronal network? This question underpins the theme of two research articles in this research topic. Using in vitro whole-cell electrophysiology from neurons in the auditory midbrain of embryonic birds, Malinowski et al. identified a link between the rapid adaptation behaviour of neurons and their intrinsic physiological properties but not their morphological classification. Phasically spiking neurons exhibiting strong adaptation with faster recovery were found to have higher thresholds, lower membrane resistance and lower membrane time-constants compared to tonically spiking neurons. These two populations exhibited few dendritic morphological differences. Conversely, neurons with different dendritic morphological structures (stellate versus elongated neurons) showed no significant difference in their intrinsic properties and spiking patterns. Quiroga et al., taking a computational approach, predicted that recurrent connections in a population of orientation selective neurons can give rise, over short time scales of a few hundred milliseconds, to attractive perceptual aftereffects instead of commonly reported repulsive aftereffects. They further employed a psychophysical paradigm, appositely designed to capture aftereffects over short time scales, in order to verify this prediction. Their results indicate how recurrent network dynamics can contribute to shape perceptual adaptation.
BibTeX
@inproceedings{adibi2021editorial,
  author = {Mehdi Adibi and Davide Zoccolan and Colin W. G. Clifford},
  title = {{Editorial: Sensory Adaptation}},
  booktitle = {Frontiers in Systems Neuroscience},
  year = {2021},
  doi = {10.3389/fnsys.2021.809000},
}
Figure