Our research
Research
One question runs through everything we do: how does the brain turn the
continuous, messy flow of experience into a memory you can hold onto?
We come at it from an unusual angle - the sense of smell. Smell is ancient, intimate,
and famously hard to put into words, which makes it a rare window onto how the brain
builds structure out of raw experience. From that starting point we follow the thread
outward: how the rhythm of breathing shapes what we take in, how an unbroken stream of
life gets carved into rememberable events, how the brain maps not only places and odors
but abstract ideas - and what happens to all of this when memory is injured, as in PTSD.
We work with both healthy volunteers and clinical populations, and we pair behavior and
physiology with neuroimaging, invasive recordings, and computational modeling.
Our research follows four intertwined threads.
Cognitive maps: from scent to concepts
The brain is a cartographer. The hippocampal–entorhinal system that lets you find your
way across a city turns out to organize far more than space - it appears to lay out
knowledge itself on map-like coordinates, where "distance" can mean the difference
between two smells, two faces, or two ideas. We begin with smell because odor space has
no built-in axes like up/down or red/green, which forces the deeper question into the
open: how does the brain build structure out of genuinely high-dimensional, non-spatial
information?
From there we ask a broader question that motivates much of the lab's current work: are
the same coordinate systems the brain evolved for physical navigation also used to map
abstract, conceptual knowledge - the way ideas, categories, and relationships relate
to one another? We measure how individuals perceive and organize odors, model the link
between a molecule's structure and how it smells, and treat each person's idiosyncratic
"smell fingerprint" as a clue to how their brain represents information more generally.
How breathing shapes the mind
Every breath is also a decision about when to let the world in. Inhalation isn't
passive housekeeping - it tunes the brain for incoming information. We've shown that
people spontaneously inhale right as a mental task begins, and that whether someone is
inhaling or exhaling at that instant changes how the brain processes the task and how
well they perform it. Because olfaction is an evolutionarily ancient sense, this
breathing–cognition link seems to reach well beyond smell, into perception, attention,
and memory.
Using lightweight wearable sensors, we log natural breathing in fine detail, in the lab
and in daily life. We're finding these patterns are strikingly individual - stable
enough to act as a respiratory "fingerprint" - and that they shift in meaningful ways
during sleep and in disease.
From continuous experience to episodic memory
Life doesn't arrive in chapters, but memory stores it as if it did. Somewhere between
the unbroken stream of perception and the tidy episodes we later recall, the brain draws
boundaries - deciding where one event ends and the next begins. We study this
event segmentation: how the brain parses ongoing, multisensory experience into
discrete events at the moment of encoding, and how the boundaries it chooses shape what we are able to remember afterward. The hippocampus sits at the center of this work as the hub where sights, sounds, and smells are bound into a single episode.
A newer strand of the lab uses AI language models to generate personalized, naturalistic
narratives —- letting us present rich, lifelike experience to participants while keeping
the experiment tightly controlled. It's a way to bring real life into the scanner without
losing the rigor an experiment needs.
Trauma, autobiographical memory, and PTSD
The same system that quietly files away an ordinary afternoon can, after trauma, refuse
to let a memory become past tense. In a study that drew coverage in the New York Times,
we found that in people with PTSD, sad autobiographical memories were represented in the hippocampus in ways that looked similar from person to person when the memories shared meaning - but traumatic memories broke that pattern, engaging the posterior cingulate cortex instead. The implication is striking: for the brain, a traumatic memory may behave less like a recollection of the past and more like an experience happening right now. We build on this with highly personalized paradigms, including individually tailored
autobiographical narratives, to trace the neural roots of emotional and traumatic memory
in both healthy participants and people living with PTSD and to point toward where
intervention might do the most good.
How we work
Across all four threads we combine:
Custom olfactory psychophysics - odor tasks that fingerprint perception at the
level of the individual.
Wearable respiratory monitoring - capturing natural breathing with unusual ease and
detail, in the lab and in everyday life.
Neuroimaging and computational modeling - fMRI analyzed with machine learning and
representational-similarity approaches to map how sensory input becomes cognitive
structure.
Multivariate pattern analysis - cross-modal decoding that links neural responses to
odors, scenes, and concepts with the memories they form.
Electrophysiology - EEG and intracranial recordings to resolve the fast timing of
memory, breathing, and event-segmentation signals.
This toolkit is reinforced by ongoing collaborations in computational psychiatry — for
example, work on how beliefs and expectations shape brain responses (Perl et al.,
Nature Mental Health, 2024)
See the full list of our work on the Publications
page, or get in touch through Contact.