Calibrating Delivery of TMS Brain Stimulation With Real-Time Neural Oscillation Patterns in the Brain To Improve Therapeutic Effectiveness
Calibrating Delivery of TMS Brain Stimulation With Real-Time Neural Oscillation Patterns in the Brain To Improve Therapeutic Effectiveness
Can non-invasive TMS brain stimulation therapy for depression and other disorders be improved so that it helps more patients? Many thousands of patients with major depressive disorder (MDD), in particular, have been helped significantly by transcranial magnetic stimulation (TMS) since it was approved by the FDA 18 years ago. BBRF Scientific Council member and two-time grantee Mark S. George, M.D., helped to pioneer the new technology and bring it to patients. TMS and its state-of-the-art variant called rTMS use a magnetic coil placed just above the scalp to induce changes in the activity of neurons of the cerebral cortex, just below the skull.
The magnetic field, which passes harmlessly through the skull and creates small local electrical currents, depolarizes neurons and stimulates them to fire—a way of inducing neural plasticity. In about one-third of patients, and by some estimates as many as half, a full course of rTMS therapy, lasting 4-6 weeks, or an accelerated protocol, lasting as few as 5 days, has the effect of reducing depression symptoms, sometimes dramatically. But the degree of effectiveness varies, and in other disorders including OCD and PTSD it remains considerably lower.
Christoph Zrenner, M.D., of the Centre for Addiction and Mental Health and the University of Toronto, is a 2023 BBRF Young Investigator grantee who is among those developing ways of enhancing the therapeutic effectiveness of rTMS and related non-invasive brain stimulation therapies. In a paper published in Nature Protocols, Dr. Zrenner with Professors Paolo Belardinelli and Ulf Ziemann advance a protocol, or set of procedures, aimed at boosting the efficacy of rTMS by targeting the delivery of its magnetic pulses with the aid of electroencephalography (EEG).
This is a fundamentally different way of targeting the magnetic pulses of TMS: the approach uses real-time EEG readouts of neural activity in the brain to calibrate the timing of TMS pulses—the delivery of each pulse or burst of pulses corresponding with a particular feature in the EEG-recorded brainwave pattern. This method is capable of changing to match changes in the EEG pattern within a couple of milliseconds to maximize the plasticity effect. (A millisecond is one-thousandth of a second.)
This “brain state-dependent” targeting of TMS pulses contrasts with TMS as used in the clinic today, which is targeted not temporally but rather spatially—typically, in depression, to a spot on the left forehead corresponding with the brain’s underlying dorsolateral prefrontal cortex (DLPFC). As demonstrated in the early 2020s by the late Dr. Nolan R. Williams and colleagues at Stanford, an fMRI functional brain scan can be used to fine-tune the precise point on the scalp to focus TMS pulses. This is designed to reflect what doctors call “inter-individual differences” among patients—the fact that the optimum spot for targeting in the DLPFC may vary by a few millimeters from one person to another.
Using fMRI brain scans to find the optimal target in each patient is one way of trying to boost effectiveness of brain stimulation by tailoring the treatment to the individual. The new protocol published by Drs. Zrenner and colleagues seeks to achieve precision-medicine individualization, in contrast, using EEG-assisted targeting.
Our brains are continually generating electrical signals, which form recurring patterns that can be detected and recorded by electrodes placed on the scalp. The easiest patterns to detect are “peaks” and “troughs” in waveforms whose separation reflects the frequency of the waves being generated by oscillating neuronal activity. They range from very slow (about one cycle per second) to very fast (100 or more cycles per second), and are associated with different functions of the brain, with sleep at the lower end and high-level cognition at the high end.
Not only do EEG patterns vary between individuals; within a single person, they can vary from moment to moment. Such variations can be quite small but may have a large effect on the efficacy of magnetic pulses delivered by TMS. The reason has to do with the wave-nature of the EEG signals and the effect of an arriving TMS pulse: the pulses, fired in rapid succession in rTMS, are of a constant intensity; but depending on the state of the patient’s brain at any given moment, those pulses may arrive at the trough of a given wave pattern, or at the peak, or somewhere in between.
As Dr. Zrenner and colleagues note, one potential way of increasing the effectiveness of TMS therapy is to calibrate the delivery of each pulse with the precise millisecond in which waves in a particular EEG wave pattern reach their trough or low point in the cycle—for example, waves in the “alpha” band (prominent during a calm and idling mental state), where the frequency is usually between 8 and 12 cycles per second. Experiments have shown that timing the delivery of rTMS pulses with the troughs of such waves will tend to maximize the degree to which the arriving pulses will increase neural plasticity, a mechanism strongly associated with the antidepressant effect.
In rTMS that is targeted spatially, it is theorized by Dr. Zrenner and colleagues that some pulses will arrive at the target when local brain activity is “out of phase,” and thus may not have the desired impact on plasticity (some pulses will inevitably arrive at the “in-phase” moment, but some or many may not). It has been suggested that out-of-phase delivery of TMS (delivery uncoordinated with neural oscillation) may account at least partly, and perhaps substantially, for non-response or weak response to TMS therapy in some patients. Hence efforts of Dr. Zrenner and colleagues to develop their new approach.
What would prevent a doctor from synchronizing the delivery of TMS with the patient’s real-time “brain state” as reflected in EEG? There are a number of reasons, mostly technical and practical. The real-time analysis of a patient’s EEG waves that the new protocol depends upon is a hard thing to pull off, and is intended, for the time being, to be a research tool. Indeed, one contribution of the newly published protocol is a detailed description of how EEG signals can be collected and analyzed on the fly, as they are being generated, and precisely situated in “brain space”, with the help of a sophisticated computer program and considerable computing power.
This reconstruction of a patient’s EEG signal “in source space,” i.e., in the 3D space of the brain, reflects neural activity over intervals as small as a millisecond. Locating the source of the targeted EEG signal in each patient’s unique brain space is based on anatomical modeling that relies upon an MRI conducted prior to the delivery of TMS therapy. So, compared with the current procedure of standard spatial DLPFC targeting, the targeting procedure that calibrates TMS with shifting EEG signals involves special expertise, extra imaging steps, and considerable computing power.
The team makes clear that their protocol is designed for research purposes, and that successful translation to the clinic will not be a matter of replicating the procedures they have published but rather adapting clinical practice to insights that they and other researchers gain from the procedure.
“Such an EEG-TMS technology enables us to make informed decisions about when to stimulate the brain with respect to ongoing oscillatory brain states,” which are “transient and ever-switching,” the researchers said.
It is in response to this particular biological property of the living brain that the EEG-TMS approach may be best suited—an approach to brain stimulation that is dynamic and, in the language of research, “closed-loop.” As opposed to the static and repeating magnetic pulses delivered by conventional rTMS, pulses delivered in a dynamic, closed-loop form of TMS would be based on a reading of the current brain rhythm or response to the prior pulse, with the next pulse adjusted in real time to match what the brain is doing at that precise millisecond.
Various proof-of-concept applications of EEG-coupled TMS have been reported or are underway in conditions ranging from MDD to chronic motor stroke. “The protocol could be reasonably tested in other brain disorders in which brain oscillations have an established physiological or pathophysiological role,” the researchers say. They propose that targeting neural activity temporally with EEG-TMS or an analogous technology could deliver improved therapeutic results in depression, OCD, addiction, Parkinson’s, stroke rehabilitation, and perhaps even dementia. The new protocol provides an avenue for researchers to conduct experiments in these and other conditions.
