Rapid task-dependent tuning of the mouse olfactory bulb

Anzhelika Koldaeva, Andreas T. Schaefer, Izumi Fukunaga

TL;DR

The mouse olfactory bulb (OB) dynamically adjusts its neural representations of odors based on the task demands, even when the odor itself remains the same. This rapid modulation enhances the discriminability of similar odors, suggesting the OB is not just a passive relay but an active processing hub that adapts to behavioral needs.

What Problem Does This Paper Address?

Our brains are constantly faced with the challenge of interpreting sensory information in a way that's useful for the current situation. What's optimal for recognizing a familiar face might not be ideal for spotting a predator. This is particularly true for olfaction, where the same odor might need to be treated differently depending on whether you're trying to identify it precisely or simply distinguish it from a similar scent. Previous research hinted that the olfactory bulb (OB), the first processing center for smell, might be involved in this adaptation, but it was unclear if these changes were rapid and directly driven by task demands, or if they were a slower consequence of learning or reward associations. This paper investigates whether the OB's output neurons can dynamically reconfigure their responses to an odor based on the immediate behavioral context, making olfactory representations more discriminable when needed.

Methods

The researchers used a clever experimental setup involving head-fixed mice trained to perform two distinct olfactory discrimination tasks: a "coarse" discrimination where odors were easily distinguishable, and a "fine" discrimination where odors were very similar. The key was that the rewarded odor (α) was the same in both tasks, allowing them to isolate the effect of task demand. They employed two-photon microscopy to image the activity of mitral and tufted cells (M/TCs) in the OB using the genetically encoded calcium indicator GCaMP6f. This allowed them to visualize neuronal responses in awake, behaving mice as they switched between tasks. They analyzed the calcium transients to quantify changes in odor response amplitude and used principal component analysis (PCA) to visualize the overall population response trajectories for different odors and tasks. To understand which neurons were responsible for the observed changes, they assessed stimulus selectivity (how well a neuron responded to one odor over another) and correlated this with task-modulated responses.

Datasets

Not discussed in the paper. The study uses data collected from experiments performed on mice.

Key Insights and Ideas

  • OB as an active, dynamic processor: The study strongly suggests the OB isn't just a passive relay station for olfactory information. It actively tunes its output based on behavioral demands.
  • Task-dependent modulation is rapid: The changes in OB responses occur almost immediately upon switching tasks, even on the very first trial after the switch. This highlights the speed at which the brain can adapt sensory processing.
  • Selective amplification for discriminability: The key insight is that the modulation isn't a uniform change in all neurons. Instead, neurons that are already tuned to differentiate between similar odors (α-selective M/TCs) show enhanced responses during the fine discrimination task. This selective amplification makes the representations of similar odors more distinct.
  • Context over association: The modulation persists even when the odor-reward association is kept constant, indicating that the immediate behavioral context, rather than learned associations, is the primary driver of these rapid changes.
  • Reversibility: The observed changes in neural representation are reversible, returning to baseline levels when the task switches back.

Key Results

  • Distinct population trajectories: PCA revealed that the population response to the same odor (α) differed significantly between coarse and fine discrimination tasks, indicating a task-dependent neural representation. These differences were reversible upon switching back to the original task.
  • Selective enhancement of α-selective neurons: During fine discrimination, α-selective M/TCs (neurons that responded more strongly to the rewarded odor α than the similar non-rewarded odor α') showed a significant increase in their response amplitude. This effect was not observed in α'-selective neurons.
  • Increased decorrelation: Removing task-modulated neurons led to increased correlation between the responses to similar odors, suggesting that these neurons are crucial for enhancing stimulus decorrelation and thus discriminability.
  • Behavioral accuracy mirrors neural changes: The recovery of behavioral accuracy after task switching closely mirrored the recovery time course of neural responses in the OB.

Limitations and Future Work

The authors acknowledge that the calcium imaging technique has limitations in terms of kinetics and sensitivity, meaning the observed 1-second response window might capture aspects of decision-making beyond pure sensory processing. They also note that while their findings point towards cortical feedback as a potential mechanism for this modulation, the exact pathways remain to be elucidated. Future work could explore the specific neural circuits (e.g., cortical feedback, neuromodulatory systems) responsible for this rapid, task-dependent tuning and investigate how this mechanism extends to other sensory modalities. Understanding these dynamic processes could offer insights into how the brain flexibly adapts to a constantly changing world.

Summary updated

Abstract

Adapting neural representation to rapidly changing behavioural demands is a key challenge for the nervous system. Here, we demonstrate that the output of the primary olfactory area of the mouse, the olfactory bulb, is already a target of dynamic and reproducible modulation. The modulation depends on the stimulus tuning of a given neuron, making olfactory responses more discriminable through selective amplification in a demand-specific way.
BibTeX
@inproceedings{koldaeva2019rapid,
  author = {Anzhelika Koldaeva and Andreas T. Schaefer and Izumi Fukunaga},
  title = {{Rapid task-dependent tuning of the mouse olfactory bulb}},
  booktitle = {eLife},
  year = {2019},
  doi = {10.7554/elife.43558},
}
Figure