Digital Olfaction in TMS Plus: Designing the Architecture of the Therapeutic Experience By A/Prof Gary Galambos

We have become increasingly sophisticated about where we stimulate the brain.

Perhaps we now need to become equally thoughtful about what a person is experiencing when we stimulate it.

In conventional transcranial magnetic stimulation (TMS), enormous attention is appropriately paid to the treatment target: MRI, neuronavigation, coil position, stimulation parameters and the brain network we are trying to influence.

At TMS Plus , we approached anatomical brain target localisation to an unprecedented level through our partnership with radiology-based VisionTMS which enables advanced precision image-guided TMS, which has been operating since 2023. It has enhanced the mental healthcare of hundreds of young adults through a collaboration with Uspace St Vincent's Private Hospital Sydney .  

But the person sitting beneath the coil is not psychologically inactive.

They are thinking. Remembering. Anticipating. Perhaps worrying. They are hearing the sounds of the treatment room and responding to the environment around them.

This is one of the ideas behind TMS Plus .

Rather than regarding that experience as background noise, we are beginning to ask whether the therapeutic environment itself can be deliberately designed.

We might think of this as creating an architecture of experience around brain stimulation.

And one of the newest elements we are exploring within that architecture is digital olfaction — scent.

From anatomical targeting to psychological-state targeting

TMS is fundamentally an anatomical intervention.

We identify a target within the brain and use a magnetic field to stimulate the relevant cortical region and associated networks.

TMS Plus retains that biological precision but adds another experimental question:

What therapeutic state is present when stimulation occurs? 

Through AI-Personalised Immersive Therapeutics (AI-PIT) , we are developing highly individualised therapeutic experiences that can be delivered during TMS.

These may incorporate clinical interviews, patient photographs, autobiographical experiences, meaningful environments and personal goals. AI can then help transform this material into immersive therapeutic scenes viewed through AR glasses during treatment.

This gives us control over more of the patient's therapeutic environment.

What do they see?

What do they hear?

What are they remembering?

What emotional or behavioural state are we trying to evoke?

What are they practising doing differently?

And now:

What do they smell? 

Why scent?

The interesting thing about smell in this context is not simply that scents can be pleasant.

Smells can become extraordinarily specific markers of people, places, experiences and autobiographical memories.

A particular smell can belong to a boat.

A garden.

A church.

A pet.

A childhood home.

Rain.

Freshly cut grass.

The ocean.

And these associations can be intensely individual.

This is why our interest in scent is very different from conventional aromatherapy.

We are not putting lavender into a treatment room because lavender is supposedly relaxing.

There is no universal “depression scent.” 

There is no “OCD scent.” 

Instead, we are asking:

What does this smell mean to this particular person? 

Reconstructing a personally meaningful world

One of our early applications involved a patient for whom sailing was personally meaningful.

His AI-generated therapeutic experience placed him back aboard his yacht.

Visually, he could see the boat and ocean. The audio helped construct the environment and therapeutic narrative.

We then began constructing an olfactory landscape around the experience.

Ocean helped represent the maritime environment.

Sandalwood contributed to the experience of the wooden boat.

Wet Earth was associated with his dog and the surrounding environment.

The important point is that none of these scents was selected because we believed it intrinsically treated depression.

They were selected because they belonged within his world.

Other patients require completely different sensory environments.

We have used combinations involving sandalwood, lavender and wet earth within an autobiographically meaningful church environment. In another case, Ocean with elements of lavender and wet earth accompanied imagery associated with an island holiday.

The scent palette follows the person.

Depression: from remembering to approaching

For depression, our AI-PIT work is particularly interested in behavioural activation, positive emotional activation, autobiographical experience, approach motivation, agency and personally meaningful goals.

Depression can progressively disconnect people from activities, relationships and environments that previously mattered to them.

An immersive therapeutic experience might therefore recreate an important place, relationship, activity or future possibility.

Digital scent potentially provides another way of making that environment personally recognisable.

But the goal isn't simply:

“Let's make someone feel happy while they receive TMS.” 

The more interesting therapeutic transition is something closer to:

“This part of my life feels inaccessible.” 

becoming:

“This still matters to me. Perhaps I can begin moving towards it again.” 

Ideally, the immersive experience connects with something achievable outside the treatment room — a real activity, relationship, place or goal that the person might begin approaching again.

OCD: the same technology with an almost opposite purpose

The application in obsessive-compulsive disorder (OCD) illustrates why this is not aromatherapy particularly well.

In depression, we might use scent to help reconstruct a positive and personally meaningful autobiographical environment.

In OCD, scent can potentially become part of an exposure environment.

Our personalised AI-ERP work creates therapeutic scenarios related to the individual's obsessions, fears and compulsive responses.

One early sequence, for example, moved through several environments:

Backyard → Fresh Cut Grass 

Bedroom → Sandalwood 

Driving environment → Creosote 

Rather than combining scents into a single emotional landscape, they can change sequentially as the exposure environment changes.

And the goal here is certainly not relaxation.

But neither is the goal simply to make someone anxious.

The therapeutic work lies in what happens after the obsession or uncertainty has been activated.

Can I experience this without checking?

Can I tolerate this without washing?

Can I resist asking for reassurance?

Can I leave something feeling “not quite right”?

Can I experience the urge to perform a compulsion without carrying it out?

Scent potentially becomes another component of the personalised exposure environment in which that new response can be practised.

Digital scent rather than a room diffuser

To explore this, we have been field-testing the Aroma Shooter® 2, developed by the Japanese digital-scent company Aromajoin .

The distinction between a digital scent device and an ordinary diffuser is important.

We don't necessarily want to fill an entire treatment room with a fragrance that remains there throughout the session.

We want to be able to direct scent towards the patient, control its timing and change it as the therapeutic environment changes.

The device contains multiple scent cartridges and can be positioned close to the patient while they view their personalised therapeutic content through AR glasses.

But the most interesting part for me isn't the hardware.

It is what becomes possible when scent becomes programmable media. 

Giving therapeutic media an olfactory soundtrack

Imagine a personalised therapeutic video with three coordinated tracks.

There is a visual track controlling what the patient sees.

There is an audio track controlling what they hear.

And now there can potentially be an olfactory track controlling what they smell.

A patient approaches the ocean.

Ocean. 

They step onto their wooden yacht.

Sandalwood. 

Their dog appears beside them.

The olfactory environment changes again.

Or an OCD exposure moves from a backyard to a bedroom and then into a car.

Fresh Cut Grass → Sandalwood → Creosote. 

The scents can potentially be programmed to change at particular moments in the therapeutic media.

I think of this as an olfactory soundtrack.

Just as music and sound effects are synchronised to a film, scent can potentially be synchronised to a therapeutic experience.

The objective is sensory coherence: what the patient sees, hears and smells belongs to the same personalised therapeutic world.

The architecture of experience

This brings us to the larger idea behind TMS Plus .

We can think of the intervention as a series of layers surrounding the patient.

MRI and neuronavigation help determine where we stimulate.

TMS provides the neuromodulatory intervention.

AI-PIT helps construct personalised therapeutic content.

AR glasses influence what the patient sees.

Audio influences what they hear.

Digital olfaction introduces what they smell.

And behavioural activation or ERP helps determine what the patient is practising psychologically and behaviourally while all of this is occurring.

The aim isn't technology for technology's sake.

It is to ask whether these elements can be deliberately organised around a therapeutic objective.

That is what I mean by the architecture of experience.

Link to Video about OINSPIRE HREC-approved research program  

Does the state of the brain during TMS matter?

Behind all of this sits an intriguing scientific hypothesis.

If we stimulate a brain network while that network is participating in a particular psychological state, might that eventually matter to the therapeutic effect?

Could anatomical targeting and psychological-state targeting interact?

There are interesting reasons from neuroscience to investigate this possibility. Goal context, motivation and behaviour can influence how neural systems represent and respond to an environment.

But moving from that observation to a clinical treatment hypothesis is a very substantial translational step.

We should not assume that creating a vivid personalised experience during TMS necessarily enhances neuromodulation.

That has to be tested .

Link to Digital Olfaction in TMS Plus Video 

Early observations are not proof

Our early clinical experience with digital olfaction has been fascinating.

Some patients have had striking subjective responses to scent when it is embedded within personally meaningful immersive experiences.

We have also seen substantial clinical improvements in some patients receiving broader multimodal TMS Plus treatment.

But these are clinical observations, not evidence that scent itself produced those improvements.

There are many things happening simultaneously.

A patient may be receiving TMS, psychological therapy, personalised AI-generated content, visual immersion, audio, scent and considerable clinician engagement.

If that patient improves, we cannot simply attribute the improvement to one component.

Likewise, a powerful emotional or autobiographical reaction to scent doesn't prove that we have engaged a particular therapeutic brain circuit.

That distinction is important.

The appropriate response to an interesting early observation isn't to declare that it works.

It is to ask a better research question.

What do we need to find out?

There are now many.

Can we reliably create the intended therapeutic state?

Does adding scent make personalised therapeutic experiences meaningfully different?

Does synchronising scent precisely with visual and auditory content matter?

Are some patients more responsive to olfactory immersion than others?

Can scent strengthen autobiographical or behavioural-activation experiences?

Can it make personalised exposure more representative of the situations encountered outside treatment?

Does any of this translate into different behaviour after the patient leaves the treatment room?

And ultimately:

Does adding digital olfaction provide measurable benefit beyond the other components of TMS Plus? 

Controlled research will be needed to answer these questions.

Link to Digital Olfaction in TMS Plus Podcast 

Towards precision multisensory interventional psychiatry

The idea that interests me most is therefore not “TMS plus a smell.” 

It is something broader.

We have spent years becoming more precise about the physical coordinates of psychiatric neuromodulation.

Perhaps the next step is to investigate whether we can also become more precise about the therapeutic experience surrounding the moment of stimulation.

What is this person seeing?

What are they hearing?

What are they remembering?

What matters to them?

What are they afraid of?

What are they learning to approach rather than avoid?

What behavioural response are they practising?

And perhaps:

What are they smelling? 

That is the emerging idea of precision multisensory interventional psychiatry — bringing anatomy, psychology, behaviour, memory and sensory experience together around the individual patient.

Digital olfaction may eventually prove to be a relatively small part of that architecture.

Or it may turn out to be considerably more important than we currently realise.

At this stage, we don't know.

And that is exactly why it is worth studying . 

---

Dr Gary Galambos  
Dr Egg’s Blogs | MindSkiller 

TMS Plus • AI-Personalised Immersive Therapeutics (AI-PIT) • Digital Olfaction • Precision Multisensory Interventional Psychiatry