Global Features of Neural Activity in the Olfactory System Form a Parallel Code That Predicts Olfactory Behavior and Perception

Rafi Haddad, Thomas G. Weiss, Rehan Khan, Boaz Nadler, Nathalie Mandairon, Moustafa Bensafi, Elad Schneidman, Noam SobelJournal of Neuroscience 2010 · DOI

Abstract: Odor identity is coded in spatiotemporal patterns of neural activity in the olfactory bulb. Here we asked whether meaningful olfactory information could also be read from the global olfactory neural population response. We applied standard statistical methods of dimensionality-reduction to neural activity from 12 previously published studies using seven different species. Four studies reported olfactory receptor activity, seven reported glomerulus activity, and one reported the activity of projection-neurons. We found two linear axes of neural population activity that accounted for more than half of the variance in neural response across species. The first axis was correlated with the total sum of odor-induced neural activity, and reflected the behavior of approach or withdrawal in animals, and odorant pleasantness in humans. The second and orthogonal axis reflected odorant toxicity across species. We conclude that in parallel with spatiotemporal pattern coding, the olfactory system can use simple global computations to read vital olfactory information from the neural population response.

Introduction

Olfactory perception is traditionally viewed as a complex "pattern-matching" problem where the brain decodes intricate spatiotemporal patterns of neural activity to identify millions of distinct scents. However, this high-dimensional complexity poses a challenge for rapid survival decisions. This study investigates whether a simpler, "parallel" code exists—a global feature of the entire neural population that provides a fast-track for critical behavioral decisions like approach or avoidance.

Methods

The researchers conducted a meta-analysis of 12 datasets across seven species, including Drosophila, zebrafish, mice, rats, and humans. They employed Principal Component Analysis (PCA) to reduce high-dimensional neural data from olfactory receptor neurons (ORNs), glomeruli (GLO), and projection neurons (PN) into its most significant axes of variation. These mathematical principal components (PCs) were then correlated with behavioral data (approach/withdrawal), human perceptual ratings (pleasantness), and toxicity metrics (oral $LD_{50}$).

Key Insights: The Dual Axes of Olfaction

The analysis revealed two primary, independent axes of neural activity that dominate the olfactory response across species.

  1. The "Pleasantness" Axis (PC1): The first principal component (PC1) accounts for the largest share of variance and correlates strongly with the total sum of neural activity. This axis effectively predicts whether an animal will approach or withdraw from an odor and maps directly to human ratings of odor pleasantness.
  2. The "Toxicity" Axis (PC2): The second axis (PC2) is mathematically orthogonal to the first and correlates with the toxicity of the odorant. This suggests the existence of a dedicated neural channel for safety assessment.
PC1 projections vs. neural activity metrics across species (Rat, Human, Mouse) and the aggregate scaled response.

Results and Robustness

The first two principal components accounted for more than 50% of the total variance in neural responses across all studied species. PC1 showed an exceptionally high correlation with the total neural sum, ranging from r = 0.73 to 0.98.

In specific model organisms, the predictive power of these global features was striking:

  • Behavioral Prediction: In Drosophila larvae, PC1 predicted approach/withdrawal behavior with a correlation of r = 0.76.
  • Human Perception: PC1 derived from mouse and human receptor responses predicted human pleasantness ratings with a correlation of r = 0.75.
  • Toxicity: PC2 significantly correlated with rat oral toxicity ($LD_{50}$) across multiple datasets (e.g., r = 0.58, p < 0.001).
PC1 projection vs. spikes per second in Drosophila (A) and Larva (B), illustrating the relationship between global neural activity and behavioral indicators.

The authors also demonstrated that this global code is highly robust. Even when sampling only 5% of available receptor data, PC1 remained highly predictable, explaining why animals can maintain basic "good vs. bad" olfactory judgments despite significant neural lesions.

Conclusion

This research suggests that the olfactory system utilizes a parallel coding scheme. While complex spatiotemporal patterns enable fine-grained odor discrimination, a simpler global code—extractable via basic summation—provides the "bottom line" for immediate survival decisions. This global code offers a robust and computationally efficient mechanism for translating sensory input into vital behavioral outputs.

Summary updated

Abstract

Odor identity is coded in spatiotemporal patterns of neural activity in the olfactory bulb. Here we asked whether meaningful olfactory information could also be read from the global olfactory neural population response. We applied standard statistical methods of dimensionality-reduction to neural activity from 12 previously published studies using seven different species. Four studies reported olfactory receptor activity, seven reported glomerulus activity, and one reported the activity of projection-neurons. We found two linear axes of neural population activity that accounted for more than half of the variance in neural response across species. The first axis was correlated with the total sum of odor-induced neural activity, and reflected the behavior of approach or withdrawal in animals, and odorant pleasantness in humans. The second and orthogonal axis reflected odorant toxicity across species. We conclude that in parallel with spatiotemporal pattern coding, the olfactory system can use simple global computations to read vital olfactory information from the neural population response.
BibTeX
@article{haddad2010global,
  author = {Rafi Haddad and Thomas G. Weiss and Rehan Khan and Boaz Nadler and Nathalie Mandairon and Moustafa Bensafi and Elad Schneidman and Noam Sobel},
  title = {{Global Features of Neural Activity in the Olfactory System Form a Parallel Code That Predicts Olfactory Behavior and Perception}},
  journal = {Journal of Neuroscience},
  year = {2010},
  doi = {10.1523/jneurosci.0398-10.2010},
}
Figure