Multimodal TMS Plus: What Happens When We Combine Brain Stimulation, AI-Personalised Therapy and Sensory Immersion?

Transcranial magnetic stimulation (TMS) is usually described as a treatment in which magnetic pulses are directed at a particular region of the brain.

That description is accurate — but it may also underestimate what becomes possible when TMS is considered not simply as an isolated procedure, but as one component of a multimodal therapeutic intervention.

An emerging question in interventional psychiatry is whether several technologies can be deliberately brought together within the same treatment course: structural MRI and image-guided targeting, different TMS protocols designed to influence cortical activity in different directions, personalised psychological interventions delivered during stimulation, AI-generated therapeutic media, and even sensory technologies capable of adding personalised olfactory information to an immersive therapeutic experience.

This is the concept behind TMS Plus 

Rather than asking only “Where should we stimulate?”, the broader question becomes: What therapeutic experience should be occurring while we stimulate — and can the different components of that experience be designed to work together?

Ingredient 1: MRI — Personalising the Map

The first ingredient is imaging.

Conventional TMS has often relied on scalp measurements or standard anatomical assumptions to estimate where stimulation should be delivered. MRI-guided approaches allow stimulation to be planned with reference to an individual’s own brain anatomy.

MRI therefore provides the map; neuronavigation helps translate that map into reproducible coil positioning during treatment.

This is an important conceptual shift. TMS becomes less about placing a coil approximately over a region of interest and more about delivering a defined stimulation protocol to a deliberately selected anatomical target.

But identifying where to stimulate is only one part of the problem.

We also need to decide how that region should be stimulated.

Ingredient 2: TMS — Upregulation and Downregulation

Different TMS protocols can produce different physiological effects.

At its simplest, the TMS Plus model can be thought of as having two broad neuromodulatory strategies.

One is upregulation: using excitatory-pattern stimulation to increase activity or responsivity within targeted networks. This is particularly relevant when treatment is trying to engage systems involved in cognitive control, motivation, positive emotion or behavioural activation.

The other is downregulation: using inhibitory-pattern stimulation when the therapeutic objective is to reduce activity or alter connectivity within networks that may be excessively active or insufficiently regulated.

The biology is considerably more complex than an “on/off” switch. Frequency, pattern, intensity, target, baseline brain state and network connectivity all matter. The effects of TMS also extend beyond the cortical tissue immediately beneath the coil because the stimulated region forms part of distributed neural networks.

Nevertheless, the upregulation/downregulation distinction provides a useful clinical framework.

It raises a further possibility: could the psychological state of the patient during stimulation also matter?

Ingredient 3: Treating the Brain While It Is Doing Something

Historically, TMS has largely been delivered as a relatively passive procedure.

The patient sits in the treatment chair while the stimulation protocol is administered.

But the brain is never actually passive.

At every moment it is processing thoughts, memories, sensations, expectations and emotional information. This raises an intriguing state-dependent question: might the activity occurring within a neural network during stimulation influence the effects of that stimulation?

This is one rationale for combining TMS with a deliberately selected psychological task.

For anxiety and obsessive-compulsive symptoms, for example, this might involve exposure and response prevention. In a different phase of treatment, the therapeutic task might instead involve positive autobiographical memories, future goals, motivation or behavioural activation.

The objective is not simply to stimulate a brain region and then provide psychotherapy at some unrelated time.

It is to investigate what happens when neuromodulation and psychologically meaningful learning occur together.

One possible mechanism is the interaction between TMS-induced changes in cortical excitability and the neural networks being recruited by the concurrent therapeutic experience. This is an attractive hypothesis, but an important distinction needs to be maintained: biological plausibility is not the same as demonstrating that simultaneous delivery is superior. That requires empirical testing.

Ingredient 4: AI-Personalised Therapeutic Media

This creates a practical problem.

If a therapeutic experience is going to be delivered repeatedly during TMS, how can it be made sufficiently specific to the individual?

Generic relaxation videos or standardised psychological content have obvious limitations.

AI-assisted media production creates another possibility.

Through a clinician-led script-development process, information about a person's therapeutic targets, memories, situations, environments, emotional associations and goals can be translated into personalised therapeutic video content.

Within the TMS Plus model, we have been developing this as AI-PIT — AI-Personalised Immersive Therapy.

Importantly, AI is not being asked to diagnose the patient or independently determine treatment.

The clinician determines the formulation, therapeutic objective and boundaries of the intervention. AI is used as a production technology: helping transform clinician-directed therapeutic material into personalised audiovisual experiences that would previously have been prohibitively difficult and expensive to create.

The resulting media can then be designed for different therapeutic purposes.

AI-ERP can recreate personally relevant uncertainty and exposure material while the individual practises resisting compulsive responses.

AI-BEAT can move in another direction — drawing upon positive autobiographical material, meaningful people and places, interests, values and future goals to support behavioural activation, motivation and positive emotional engagement.

And AI-REFLECT can help integrate and consolidate the experience, turning observations from treatment into a coherent therapeutic narrative.

Ingredient 5: Adding Smell to the Therapeutic Environment

Visual and auditory immersion are only part of human experience.

Smell is particularly interesting because olfactory processing has unusually direct relationships with neural systems involved in emotion and memory.

That raises a fascinating therapeutic possibility.

What if the sensory environment accompanying a personalised therapeutic video could also include programmed olfactory stimulation?

Recent TMS Plus courses have begun exploring a programmable scent device capable of releasing different scents at selected points according to the content of a personalised therapeutic video.

A scene involving an outdoor environment, for example, can potentially be accompanied by an appropriate environmental scent. Another scene can use an entirely different olfactory stimulus.

The objective is not simply to make a video more entertaining or realistic.

It is to investigate whether adding congruent sensory information can make the therapeutic experience more embodied and contextually salient.

There are several biologically plausible reasons why this might be interesting. Olfactory pathways have close anatomical relationships with limbic structures involved in memory and emotional salience, including the amygdala and hippocampal systems. Smell is also particularly powerful in evoking autobiographical memories and emotional states.

This makes personalised olfactory stimulation an intriguing candidate for both exposure-based interventions and positive autobiographical interventions.

But the distinction between hypothesis and evidence is critical.

We are not suggesting that scent itself is an established treatment for OCD, depression or other psychiatric disorders. Rather, programmable olfactory stimulation is being explored as another potential ingredient in creating a richer, more personalised therapeutic state during a multimodal intervention.

Putting the Ingredients Together

The interesting part of TMS Plus is therefore not any single technology.

It is the possibility of orchestrating them.

An individual course might bring together:

MRI → precision targeting

TMS → modulation of selected neural networks

Psychotherapy → a defined learning task

AI → personalised therapeutic content

Visual and auditory immersion → engagement of the relevant psychological context

Programmable scent → an additional personalised sensory dimension

The treatment can then change over time.

One phase might concentrate on downregulating pathological threat or arousal processes while undertaking exposure-based work.

Another might increasingly emphasise excitatory or upregulatory approaches alongside positive autobiographical material, behavioural activation and personally meaningful goals.

In this model, the course is not necessarily one protocol repeated many times.

It can become a sequence of deliberately designed brain states and therapeutic experiences.

Why Might Combining These Technologies Matter?

The hypothesis underlying multimodal TMS is that neuromodulation may be more therapeutically useful when we think not only about where and how the brain is stimulated, but also about what the brain is doing at the time.

TMS can alter cortical excitability and network activity.

Psychological interventions can activate particular cognitive, emotional and behavioural networks.

Personalised audiovisual material may increase the relevance and salience of those experiences.

Olfactory stimulation may add another route into memory and emotional circuitry.

MRI and neuronavigation allow these interventions to occur while stimulation is being delivered to a deliberately selected anatomical target.

The intriguing scientific question is whether these ingredients simply coexist — or whether, under the right circumstances, they interact.

Could stimulation make a network more receptive to therapeutic learning?

Could a carefully chosen psychological task influence which networks are active during that period?

Could personalised multisensory information make that learning more salient?

Could an intervention initially focused on reducing pathological threat processing later transition toward strengthening reward, motivation and goal-directed behaviour?

These remain research questions rather than settled conclusions.

But they are testable questions.

From TMS to Multimodal Interventional Psychiatry 

Perhaps the larger implication is that the future of interventional psychiatry may not be defined by increasingly sophisticated stimulation devices alone.

The next step may be learning how to combine technologies intelligently.

MRI can tell us more about where we are intervening.

TMS gives us a means of influencing neural activity.

Psychological science tells us what needs to be learned.

AI makes highly personalised therapeutic media increasingly practical to produce.

Immersive audiovisual and olfactory technologies allow that therapeutic information to be delivered through multiple sensory channels.

The opportunity is to bring those capabilities together while preserving something essential: the clinician remains responsible for the formulation, therapeutic strategy and treatment decisions.

Technology supplies new instruments.

The clinical task is learning how — and when — to conduct them together.

That is the emerging idea behind multimodal TMS Plus.