The team focused on the olfactory system because it offers an unusually clean experimental window into memory. Odors are high-dimensional chemical stimuli, yet the brain rapidly compresses them into compact internal representations. When an animal learns that a particular odor predicts a reward or a punishment, the neural code for that odor changes. The researchers set out to determine exactly how those changes unfold at the level of entire populations rather than single cells.

Using large-scale recordings from mice as the animals learned to associate specific odors with outcomes, the investigators tracked activity in key nodes of the olfactory memory circuit, including the piriform cortex and connected structures. Dimensionality reduction techniques revealed that odor representations occupy smooth, low-dimensional manifolds embedded in the high-dimensional space of population activity. Learning did not simply add noise or scatter the responses; instead, it systematically reorganized the geometry of these manifolds.

Specifically, as animals became better at discriminating rewarded from unrewarded odors, the manifolds corresponding to different odor categories moved apart, increasing the margin between them. This geometric separation mirrors the objective functions used in machine learning classifiers, which seek to maximize the distance between classes. In other words, the biological network appeared to solve an optimization problem: reshape its internal representation space so that behaviorally relevant distinctions become as easy as possible to read out.

The study also examined how this optimization is implemented mechanistically. Plasticity at synapses within the olfactory cortical network provides the natural substrate for manifold reshaping. By adjusting connection strengths in a way that depends on task demands, the circuit can rotate, stretch, and translate the manifolds on which odor memories reside. Computational models in the paper demonstrated that a simple learning rule acting on recurrent connections is sufficient to reproduce the observed geometric changes and the accompanying improvements in behavioral performance.

One of the most striking findings is that manifold reorganization predicted behavior on a trial-by-trial basis. In sessions where the geometric separation between odor categories was larger, animals performed the discrimination task with greater accuracy. This tight coupling between representational geometry and action strengthens the argument that the manifold is not an epiphenomenon of recording analysis but a functional unit of memory itself, something the brain actively constructs and maintains.

The work also addresses a long-standing puzzle in memory research: why memories remain stable even as individual neurons change their firing properties. If a memory were stored in the activity of particular cells, drift in those cells should degrade the memory. But if the memory is stored in the shape and position of a manifold, the system can tolerate turnover and drift at the single-cell level as long as the global geometry is preserved. The olfactory network appears to exploit exactly this robustness, stabilizing the manifold while individual neurons come and go from the active ensemble.

These results resonate with a broader theoretical movement in neuroscience that treats population activity through the lens of geometry and dynamics. Rather than decoding the activity of single neurons, this framework asks how entire trajectories and subspaces of activity support computation. The new findings provide some of the clearest evidence yet that learning sculpts these subspaces deliberately, and that the rules governing this sculpting can be described with the same mathematical language used in machine learning.

The implications reach beyond olfaction. Manifold optimization may be a general principle of representational learning in the brain, applying to motor skills, decision-making, and perhaps even higher cognitive functions. If so, therapeutic strategies for disorders of memory and perception might one day aim not at individual synapses but at restoring healthy manifold geometry in malfunctioning circuits. The study also suggests that artificial neural networks, which already borrow heavily from brain-inspired principles, could benefit from architectures that explicitly optimize manifold structure the way biological networks appear to do.

Future experiments will aim to identify the precise plasticity mechanisms and neuromodulatory signals that guide manifold optimization, and to test whether similar geometric learning rules operate in other sensory and memory systems. For now, the study offers a compelling synthesis: memory is not a static snapshot written into cells, but a dynamically optimized shape in the brain’s representational space, continuously refined by experience until the world’s important distinctions stand out in sharp relief.

The piriform cortex, the largest olfactory cortical area, occupies a distinctive position among sensory cortices. Unlike primary visual or auditory regions, it receives direct input from the olfactory bulb without an intervening thalamic relay, and its recurrent collateral connections are extraordinarily dense. Pyramidal cells in this structure broadcast axonal branches widely across the network, meaning that any learning rule acting at these recurrent synapses has access to an almost associative memory-like architecture. This anatomical arrangement has long suggested that the piriform cortex functions as a pattern completion and pattern separation device, and the manifold optimization account fits naturally within that tradition, extending it from single-cell response changes to the collective geometry of ensembles.

The statistical structure of odor space itself provides important context for why the olfactory system might be particularly suited to geometric reorganization. Natural odorants are mixtures of many volatile molecules, and the relationships between odorants are smooth: chemically similar molecules tend to smell alike, and perceptual similarity decays gradually with molecular distance. The early stages of the olfactory pathway, from receptors in the nose through the glomerular map of the olfactory bulb, already impose a dimensionality reduction on this chemical space. What the new findings add is evidence that later, experience-dependent stages continue this compression but do so adaptively, stretching the dimensions that matter for current behavioral goals while letting irrelevant dimensions collapse.

Classical work on olfactory learning in rodents emphasized changes in single-neuron selectivity, with individual piriform neurons broadening or sharpening their tuning after conditioning. Those observations were sometimes difficult to reconcile, because different cells appeared to change in inconsistent directions. The manifold perspective resolves this apparent inconsistency: heterogeneous single-cell changes can produce a coherent geometric shift if they collectively move or reshape the ensemble’s low-dimensional subspace. A neuron increasing its responses to one odor while a neighbor decreases its responses may look contradictory at the single-cell level, yet both changes can contribute to expanding the distance between odor-category manifolds, exactly what improved discrimination requires.

The distinction between representational drift and manifold stability also connects to an active debate about how the brain balances flexibility and permanence. Longitudinal recordings in several systems have documented that the identities of neurons participating in a code can change over days and weeks, even when behavioral performance is fully preserved. Explanations of this drift have ranged from passive turnover to active consolidation processes. The finding that geometric structure is preserved while its cellular substrate rotates suggests that the relevant invariant for memory is relational rather than compositional: what matters is how representations sit relative to one another, not which particular neurons carry them. This reframing gives theorists a concrete target for models of memory maintenance over long timescales.

Connections to machine learning deepen the significance of the results. Classification algorithms such as support vector machines explicitly maximize margins between categories, and deep networks trained with error-driven rules are known to linearize their internal representations as performance improves, spreading class clusters apart in hidden-layer activity spaces. The apparent convergence between these engineered objectives and the geometry observed in olfactory cortex suggests that margin maximization may be a general computational principle that biological networks arrived at independently. It also raises the question of whether the brain’s learning rules approximate gradient-based optimization over a representational objective, or whether local synaptic plasticity merely produces margin expansion as an emergent byproduct. The computational models in the study, which reproduce the geometric changes with simple recurrent plasticity, favor the latter possibility, indicating that no explicit global error signal is required.

Methodologically, the study illustrates the growing power of combining large-scale electrophysiology or imaging with tools from applied mathematics. Dimensionality reduction methods, including approaches that preserve local neighborhood structure and those that track low-dimensional trajectories over time, allow researchers to ask quantitative questions about representational geometry that were previously unanswerable. Trial-by-trial alignment between geometric measures and behavior exemplifies a broader trend: rather than treating population analyses as descriptive, investigators now use them to generate predictions that can be tested against the animal’s choices on individual trials, tightening the link between neural data and computation.

Several questions remain open. Whether manifold optimization operates during passive exposure to odors or requires explicit reinforcement is unresolved, as is the role of top-down signals from prefrontal or hippocampal structures that inform the olfactory cortex about task context. The timescale of geometric consolidation, and whether reshaped manifolds persist during sleep-related replay, would clarify how learning is stabilized. Answering these questions will require the same combination of longitudinal population recording, behavioral quantification, and geometric analysis demonstrated here, applied across circuits and species. The study thus serves both as an empirical advance for olfactory neuroscience and as a methodological template for testing whether representational optimization is a universal grammar of learned neural computation.

Subject of Research: Representational learning through geometric optimization of neural manifolds in the olfactory memory network

Article Title: Representational learning by optimization of neural manifolds in an olfactory memory network

Article References: Hu, B., Temiz, N. Z., Chou, C.-N., Rupprecht, P., Meissner-Bernard, C., Titze, B., Chung, S., & Friedrich, R. W. (2026). Representational learning by optimization of neural manifolds in an olfactory memory network. Nature Neuroscience. https://doi.org/10.1038/s41593-026-02429-3

Image Credits: AI Generated

DOI: 10.1038/s41593-026-02429-3

Keywords: neuroscience, olfaction, neural manifolds, memory, representational learning, piriform cortex, neural plasticity, population coding, dimensionality reduction, behavioral learning, machine learning, Nature Neuroscience

Brain-computer interfaces represent one of the most promising frontiers in neuroscience, offering the potential to restore lost functions and augment human capabilities by translating neural signals into commands for external devices. However, despite rapid advances in hardware and signal processing, widespread implementation of BCIs has been hindered by the sluggish and inconsistent process by which users learn to control these systems. Each individual exhibits unique neural signatures and adaptation profiles, making it challenging to develop universally effective training paradigms.

Addressing this critical bottleneck, the research team hypothesized that the brain’s spontaneous activity patterns are constrained within a low-dimensional intrinsic manifold, a concept borrowed from modern data diffusion techniques. This geometric representation encapsulates the principal directions along which neural activity naturally varies, thereby defining a latent structure within high-dimensional brain data. The core premise of the study was that BCI mappings aligned with the intrinsic manifold would facilitate more efficient neural reconfiguration during learning, whereas mappings orthogonal to this manifold would impede control acquisition.

To test this idea, participants were engaged in an immersive real-time functional magnetic resonance imaging (fMRI) experiment requiring them to control an avatar within a video game environment solely by modulating activity in brain regions implicated in spatial navigation. Unlike conventional BCIs that commonly rely on motor or sensory cortex signals, targeting higher-order cognitive brain areas posed an added challenge but also heightened ecological validity, reflecting complex, abstract neural computations. Critically, this setup allowed the researchers to experimentally manipulate the relationship between the recorded brain activity and the avatar’s movement along specific geometric axes derived from the intrinsic manifold.

The experimental design incorporated systematic perturbations to the brain-to-avatar decoder mappings. In some conditions, these mappings corresponded to directions of significant variance on the intrinsic manifold, essentially exploiting the brain’s natural activity corridors. In other conditions, the mappings were deliberately chosen to lie outside or orthogonal to this manifold, representing unnatural or difficult-to-achieve patterns of activation. The researchers then quantified participants’ ability to learn control, monitoring both behavioral performance and neural adaptation over time.

Findings were unequivocal: when the decoder mappings adhered to the intrinsic manifold geometry, participants rapidly learned to exert precise control over the avatar. This success was underpinned by realignments of neural activity along the manifold’s principal directions, indicating that learning leveraged existing neural pathways rather than forging entirely novel ones. Conversely, when decoder mappings defied the manifold constraints, participants struggled significantly, rarely achieving mastery despite equivalent training durations. This stark contrast underscores the manifold’s role as a fundamental neural scaffold that shapes learning trajectories and constrains plasticity.

Delving deeper into neural dynamics, the study revealed that successful BCI learning involved not just superficial modulation of brain activity but a profound reorganization of intrinsic manifold geometry itself. Participants’ neural activity patterns became more tightly clustered around task-relevant directions, suggesting an adaptive refinement of network-level functional architecture. This phenomenon implies that the brain optimizes information processing efficiency during BCI skill acquisition by reinforcing pre-existing geometrical frameworks.

Beyond its immediate experimental insights, this research holds transformative implications for the design of next-generation neurotechnologies. Incorporating intrinsic manifold geometry into BCI decoder construction could lead to personalized interfaces that adapt to users’ unique neural landscapes, reducing training times and enhancing overall robustness. Furthermore, this geometrically grounded approach offers a unifying theoretical framework to interpret individual differences in learning rates and aptitude for neurofeedback tasks.

Importantly, the methodology employed—leveraging advanced data diffusion algorithms to extract latent neural manifolds—exemplifies the power of contemporary computational neuroscience tools to bridge the gap between abstract mathematical constructs and practical clinical applications. By marrying sophisticated dimensionality reduction techniques with real-time neuroimaging feedback, the study demonstrates how cutting-edge analytics can illuminate the hidden structures within brain function that govern learning plasticity.

The choice of spatial navigation brain regions as a model system also highlights the generalizability of the intrinsic manifold concept across varied cognitive domains. While sensorimotor BCIs remain dominant, expanding focus to higher cognitive areas opens avenues for controlling complex, multidimensional mental states, potentially enabling revolutionary applications in communication, rehabilitation, and augmented cognition. Intrinsic manifold-guided training could become central to unlocking these capabilities.

Moreover, this work prompts a reevaluation of what constitutes effective neurofeedback paradigms. Rather than imposing arbitrary mappings that users must painstakingly override, grounding BCI design in the brain’s native activity geometry respects biological constraints and aligns with inherent learning mechanisms. This paradigm shift aligns with emerging perspectives advocating for harmonization between brain-machine interfaces and intrinsic neural processes for maximum synergy.

The scalability of this approach also bears promising prospects. While fMRI was the primary tool for capturing large-scale brain activity with high spatial resolution in this study, future integration with more portable modalities like EEG or functional near-infrared spectroscopy could enable manifold-informed BCIs in real-world settings. Such advances would democratize access to personalized neurotechnologies, impacting a broad spectrum of users from clinical populations to healthy individuals seeking cognitive enhancement.

In sum, the discovery that human learning of noninvasive BCIs is robustly mediated by brain manifold geometry offers an elegant solution to a vexing problem in neuroscientific engineering. By uncovering the geometric constraints shaping neural plasticity and demonstrating their utility in guiding decoder adaptation, the investigation lays a foundational principle for advancing human-machine symbiosis. As BCIs evolve toward clinical maturity and everyday use, embedding intrinsic manifold considerations will be instrumental in realizing their full therapeutic and augmentation potential.

This landmark study intersects diverse fields including cognitive neuroscience, machine learning, applied mathematics, and neuroengineering, representing a paradigm-defining leap in our understanding of how complex cognitive tasks can be harnessed and enhanced through technology. By transforming the way we conceptualize brain control, it paves the way for future breakthroughs where neural function and artificial systems are embedded seamlessly within a shared geometric framework. The implications for rehabilitation, medicine, and human-computer interaction are profound, heralding a new era where brain interfaces are not just tools, but extensions of intrinsic neural architecture.

As the field continues to unravel the manifold underpinnings of cognition and learning, it becomes increasingly evident that mastering the geometry of brain activity is key to unlocking human potential in a digitally integrated world. This seminal contribution by Busch, Fincke, Lajoie, and colleagues thus marks a critical inflection point, bridging theoretical neuroscience and practical innovation at unprecedented depth. The promise of accelerated, intuitive BCI learning finally approaches reality, signaling transformative possibilities for millions globally.

Subject of Research:
Human learning of noninvasive brain-computer interfaces guided by intrinsic manifold geometry of neural activity.

Article Title:
Human learning of noninvasive brain–computer interfaces via manifold geometry

Article References:
Busch, E.L., Fincke, E.C., Lajoie, G. et al. Human learning of noninvasive brain–computer interfaces via manifold geometry. Nat Neurosci (2026). https://doi.org/10.1038/s41593-026-02311-2

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41593-026-02311-2