Share This Page
Amit Etkin, M.D., Ph.D.
Associate Professor, Psychiatry & Behavioral Science
Investigator, VA Sierra-Pacific Mental Illness Research Education and Clinical Center
Dana Foundation Grantee: 2011-2016
Your laboratory has been investigating the use of repetitive transcranial magnetic stimulation (rTMS) in combination with whole-brain EEG and functional MRI (fMRI) to treat depression and to help unravel its underlying brain circuitry. First, why has depression been so difficult to treat?
There are many reasons, but I think you have to start with the fact that we don’t really know what the disorder is about. Different people report different symptoms, which may reflect biologically distinct conditions. It stands to reason that the nature of the intervention can’t possibly be the same for all those different forms of what we call depression. So partly, we need to better understand the biology of depression.
Related to that is the fact that we’ve never developed a clear idea of what we’re targeting. All of our current interventions have been discovered by serendipity. To this day we don’t really know how antidepressants work. We know a lot of things change in the brain when you take an antidepressant, but we don’t really know what is causing those changes because we don’t really know what’s wrong in the first place.
How can science move beyond this quagmire?
In an ideal world, we could design an intervention that we understood, that we could control, and that could tell us if we’re on the right track in changing what we want to change. That’s why we decided to do this study.
With medication, there is no way to alter how the medication works to better fit some particular brain target. But with targeted brain stimulation, you have an infinite amount of flexibility. For example, where do I stimulate in the brain? Does it matter if I stimulate the left side or the right side? Frontally or in the parietal cortex? Which part of the frontal cortex? Does it matter how I do the stimulation? Which frequencies or pulse patterns produce the downstream effects I’m looking for?
With this combined TMS-fMRI approach, we can produce a physiological readout that tells me if I’m hitting my target and am on the right track or not. If I need to, I can immediately devise a different protocol that engages that target better and assess whether the readout changed. Therapeutically, it puts us in a much better place than where we are now.
How is this work helping to elucidate the systems neurobiology of depression?
We need to have a handle. In this case, the critical handle is causality. One reason depression and other psychiatric conditions have been hard to understand based on brain imaging is that you inevitably find differences in patients’ brains. You might find an overwhelming number of differences all happening at the same time, and it’s impossible to know what’s related to the clinical state vs. what might be compensatory changes.
Having the tool of TMS allows me to ask what happens when I stimulate this part of the brain, or when I turn this part up or this part down. It‘s a brain-systems level intervention that carries with it the really important component of causality.
Your lab at Stanford is one of only a handful worldwide who are using TMS and fMRI concurrently. What are the benefits of combining these two?
The big advantage is causality. It gets us out of the chicken-or-egg problem where we see a lot of things change in the brain but we don’t really know why. It also allows us to go very quickly from assessing brain circuits to treating brain circuits, because the tool that we’re using to understand what’s working and what’s not can then be applied therapeutically to induce long-lasting change in a circuit. You can’t do that with imaging alone.
Some brain structures can simply not be assessed without fMRI. The amygdala, which we think is very important for a range of emotional and behavioral disorders, is one example: you can’t assess amygdala activity with EEG, and you can’t really know if you have affected it by looking only at behavior. The ability to image the amygdala and other deep brain structures in response to discrete TMS stimuli gives you a huge advantage in terms of finding new targets and new methods to modulate these regions.
How does this work advance the treatment of depression?
By stimulating brain activity and assessing circuit-level changes as they happen, we can garner important insight into what is wrong in depression and how to fix it in an optimized, personalized matter.
I’ll give you one concrete example: It matters whether stimulation is done to an area in the patient’s brain that is abnormal or normal. For any treatment in any psychiatric disorder, we don’t actually know whether the goal of treatment is to normalize abnormal brain activity or to engage compensatory circuitry. It’s a fundamental question that we cannot answer without a direct tool for manipulating brain systems and assessing the effects.
You’ve also applied EEG to your studies, and have found a signature pattern of brain activity that tells you whether the person is likely to respond to a particular regimen of TMS. How does that inform treatment?
That EEG signature not only tells us who is likely to respond, it tells us a change happened and the magnitude of that change. Those three things combined give us a platform for optimizing rTMS for depression. If we can replicate our initial result—and we’re trying to do that now—we should be able to tailor stimulation to any given person based on their EEG readout. We’ll be able to know how well we’re doing and what kind of protocol to use.
It gives us a target, a signal to chase. Then you can try different protocols and optimize your treatment fairly quickly until you find the signal you’re chasing.
How does this research advance the nascent field of clinical TMS?
The bigger picture and long-term significance is that we can start thinking about rTMS a lot more powerfully, as a personalized tool that we can optimize. Unlike any other tool in psychiatry, we can treat to a target rather than, for example, blindly elevating the dose of an antidepressant medication without knowing what brain system we’re trying to affect. We can even think of something called closed-loop TMS, where we can read out the effects as the procedure is done and adjust parameters as necessary based on real-time assessment of how neural circuits are responding.
All of these things would radically shift how we use rTMS in the clinic for depression and open up a completely new window for using TMS to target other systems. We talk a lot about personalized medicine within psychiatry and medicine at large. This is personalized medicine in essence because we can optimize the therapy as we go vs. just finding a match between an existing intervention and a profile of symptoms. I’m hoping this becomes the beginning of a shift in how we approach depression and how we optimize and personalize treatment for depression.
Has TMS really caught on as a treatment in psychiatric disorders?
Sort of. It’s used in the clinic, and is paid for by insurance companies fairly routinely in most parts of the country. Yet, psychiatrists aren’t really used to doing procedures, and TMS is a procedure. They’re not used to touching their patients, and they’re not used to working with brain circuitry, so it takes a bit of a shift in mindset within psychiatry. Also, the experience of how to do TMS is fairly limited in the field, and it has been limited somewhat by the business model of one of the leading device makers, which has decreased the flexibility of clinicians to innovate. Consequently, relatively few patients have access to TMS.
Do you see potential outside of depression in other psychiatric disorders for this method that you are developing?
Yes, absolutely. The bulk of the published research so far has been in depression, but there is absolutely no reason why you couldn’t use TMS in a proper and flexible manner in other conditions. We’re using TMS/fMRI in patients with post-traumatic stress disorder and in stroke recovery. Others are investigating it in autism and bipolar disorder.
We’re now considering other brain targets, other EEG or fMRI signals, and other plasticity protocols aimed at inducing long-term change in the brain. You can move beyond disorders per se and think about how to use these principles to characterize an individual’s brain, to say for example, “this circuit is intact and this one is not,” and then target the stimulation protocol accordingly. It may be that in some people you want to decrease a particular target signal but in others you want to increase it.
I think one should see this as the germinal phase of applying TMS to psychiatric disorders. We’ve found a signal that we can use to understand how the brain works and how our interventions can be optimized. We see this is as just the tip of the iceberg.