By: Ted Williams, M.D.
How do we get higher remission rates with rTMS?
One answer: Target the right regions, which are different in different patients.
I have a question for my fellow TMS practitioners: when you treat a patient with TMS who has been depressed for years and “tried everything,” and responds so well to treatment that by the end their face is beaming? They’ll say things like, “You’ve given me my life back,” or “I never thought I would feel this way again.” It’s those patients that make TMS the most rewarding treatment modality in psychiatry. But on the other hand, how deep the disappointment when it doesn’t work! Zip. Zero. Nada. No improvement whatsoever. The patient had put their last hope in TMS, and it did not work. As a clinician, when this happens, I wonder, “Did I miss the target?” or “Why didn’t it work? Could I have done anything differently?” And why does it matter to us? It matters because our patients are valuable human beings and their lives matter.
Remission. This is the issue Jonathan Downar addressed at the Clinical TMS Society Annual Meeting in San Diego this year. What can we do to improve the rates of remission? Even with MRI-guided treatment, many patients do not respond to TMS. Why?
Downar asks, “When we say we get 33% remission rate in controlled studies for depression with rTMS, do we really mean that we've got a technique that works in a 100% of people and succeeds 30% of the time? Or is it more likely that we've got a technique that's 80% successful in subgroup of patients that are, say, 40% of the patients who come through our door?”
So how long will rTMS remain a one-trick pony of treatment over the DLPFC? Can we find neural subtypes of depression that would better inform us as to where to point our TMS coils? For those patients who do not respond to TMS as it is currently practiced, over the DLPFC, if we knew where to target patients with other subtypes of depression, could we get better results? Can we get more patients into remission?
TMS researchers such as Jonathan Downar are making progress in identifying subtypes of depression based on functional connectivity of networks associated with symptom domains that occur across the spectrum of diagnoses in the DSM. If we can identify the pathological networks, and target their nodes on the surface of the cortex, we can make neuroanatomical-clinical formulations of our patients and more effectively treat patients with TMS.
But how do we know where to treat each patient?
Evidence is building that patients with Major Depressive Disorder as diagnosed in the DSM-5 are a heterogeneous group, all with commonalities of symptoms outlined in the DSM criteria. However, the symptoms of each patient and the way each patient presents, to some extent, their unique underlying pathology in the funcitonal connectivity of their brain networks. Subtypes of depression have different network pathologies. In order to individualize treatment of patients, we must identify the subtype of depression that each patient has, according to which netowork is not funcitoning properly, and then stimulate the pathological network, rather than our current method of stimulating everybody over the DLPFC.
EEG studies have identified 3 subtypes of depression
EEG patterns may be one way to begin subtyping patients with depression. For example, Webb and colleagues have identified three subtypes of depression
based on EEG patterns and symptoms that correlate to the three networks identified to be associated with depression. More about the three networks later.
How does TMS work, on the network level?
Answer: TMS strengthens or weakens synapses in cortico-striatal-thalamic loops.
How does TMS work? On the network level, TMS works by strengthening or weakening synapses in the circuits of cortico-striatal-thalamic loops. Sorry. That’s what they’re called. See the companion article, “Digging Deeper” to see a more complete explanation. Think of the images you have seen of TMS stimulation in which stimulation at the local site has distant effects. The brain is complicated, and processing the signals of circuits is what the brain does to carry out its functions. Typically, in rTMS, we stimulate the DLPFC, which is connected to the striatum & basal ganglia, which is connected to the thalamus, which projects back up to the DLPFC, a loop. In some patients, the network is “discombobulated,” a Downar description that is apt.
Which Cortico-Striatal-Thalamic loops are involved in Depression? Where is the lesion?
Downar affectionately calls the 7 network families and 17 sub-networks that have been identified “the Periodic Table of the Brain.” A growing understanding of these networks helps us understand what constitutes
a “sick network,” and will likely contribute to the NIH effort, RDoC (Research Domain Criteria), for understanding brain function and illness in a more biologically-based, rather than the DSM’s phenomenonologically-based, manner. Of the seven major networks, only three appear to be of particular interest to psychiatrists. Following are the three networks.
1. The Salience Network
The most important network is the Salience Network (our current target in rTMS), which enables cognitive control. It is a “thought interrupter.” When explaining it to my patients, I call it the “blow-off-er.” It’s what enables a spouse, for example to blow off the negative comments of the other, keep things in perspective, and not dwell on them. If the salience network is down, then one can't shut down their negative intrusive thoughts.
TMS of the Salience network causes growth of gray matter. Remarkable!
Put another way: TMS takes your brain to physical therapy
Although the point could be argued, the most important news coming out of this year’s annual meeting is the discovery that, not only does TMS cause release of dopamine in neural circuits, it grows gray matter. Remarkable. When I described this to one of my patients, she said, “Oh, it’s kinda like TMS takes my brain to physical therapy.” Well said.
“So we think based on emerging evidence, the one of the important things rTMS does is if you put it over the same place of a person's brain such as the dorsolateral or the dorsomedial PFC, and you do 30 days of stimulation, you're actually changing the gray matter in the cortico-striatal-thalamic loop circuits that drive, i.e., initiate, maintain, and terminate activity, in the piece of cortex that you're stimulating.” So whatever loop we're going after, that's what we're doing. And that's helpful because there are different loop circuits that go with different symptom clusters in this heterogeneous entity that we call major depression.
The Salience Network involves many of the major psychiatric disorders, not just depression
"And likewise, [the Salience Network impacts] not only depression – but also bipolar disorder, schizophrenia, OCD, eating disorders, substance use disorders. All these different things that we see coming through our clinic and we wondered, “could we use an rTMS coil to help this person?”
So the question is, which loop circuits should we be targeting to hit the different dimensions of psychiatric illness?”
What does a patient with a Salience Network deficiency look like?
When describing a patient who has difficulty with cognitive control from a Salience Network deficiency, Downar made us laugh:
“Does anyone else have in their practice a patient who sort of has a little bit of everything in the DSM? So bipolar NOS slash/ but then you can't really get a full hypomanic episode slash/ ADHD, binge eating disorder. I used to purge when I was younger now I've stopped. I used to cut myself now I don't do that anymore. I've got PTSD but it's mostly about my dog dying and then tendency of when I drink I tend to binge drink. Shows up 20 minutes late to the appointment because they couldn't get disengaged from whatever they were doing before. They photocopied the form backwards and they filled out everything in the wrong order and there's this kind of-- and then you talk and then the history takes forever because their thought form is completely circumstantial and over-inclusive. Does anyone have [laughter in the audience] patients like this?”
An Observation: Are we not seeing a convergence of modalities of treatment?
Hmmm. TMS grows gray matter. Antidepressants increase BDNF and grow gray matter. Wait, don’t I remember seeing articles about mindfulness being associated with gray matter change? With this information, are we not seeing a confluence of different modalities of treatment targeting the same pathology?
2. The Reward Network
The Reward Network enables motivation and pleasure.
The next network to consider is the Reward Network that enables motivation and pleasure. If the Reward network is down, then one can't motivate oneself to do things. If you think about it, simply brushing your teeth requires activating the Reward network, as does deriving pleasure from any activity. Reward network deficiency is lack of reward. But the Reward network is sometimes not deficient, but corrupted. It can be “hijacked,” meaning pathological incentives have become hyperconnected in the network – incentives such as addictions and drives for thinness. Since the Reward Network is anticorrelated to the Salience network, treating the Salience Network (rTMS over the DLPFC/DMPFC) may sometimes address Reward network issues by disrupting the Reward network when it is pathologically incentivized.
3. The Non-Reward Network
The Non-reward network is involved in the generation and regulation of negative affect
A relatively new network to be identified that holds significant promise for helping patients who do not respond to TMS over the DLPFC/DMPFC (Salience Network) is the Non-Reward Network. What does this mean, “Non-reward?” Non-reward functions refer to the generation and regulation of negative affect, e.g. loss, preoccupation, obsessiveness, hypotheticals and “fictive error” (discrepancy between actual outcomes and ‘could have been’ outcomes). It is also involved in “reversal learning” so when the Non-reward network is not functioning properly, it is difficult to engage in reversal learning, and causes a tendency toward obsessiveness. Inhibitory treatment to this area is helpful because the network is overly active and creating a drive or incentive toward negative thoughts, emotions, and anxieties.
What does a patient with a Non-Reward Network that is hyperconnected look like?
Downar describes these patients as ones who are “locked-on” to negative thoughts. They are frequently anxious, obsessive, and anhedonic. They are anhedonic because the Non-reward network has overpowered the Reward Network. Treating it with inhibitory TMS means inhibiting the hyperconnected network – weakening the synapses in that circuit.
Clinical implications for rTMS of the Three-network theory of depression
What are the clinical implications of the emerging evidence that there are three networks that have been associated with depression?
The most obvious implication of the network theory is that standard rTMS, over the right DLPFC, is likely to unhelpful, or only partially helpful, to those patients who do not have Salience network pathology as their primary problem. Patients can be depressed and still have an intact, functioning Salience network. When we treat those patients with rTMS, we are treating a problem the patient doesn’t have.
Another reasonable conclusion from the theory is that since the Salience network is involved in several psychiatric disorders, targeting it could be helpful in other disorders.
Patients who have pathology in more than one network: when dumb may be better than smart.
It is reasonable to assume, given the heterogeneous nature of Major Depressive Disorder, that a patient’s pathology may sometimes be in more than one network. It may not be possible, through symptoms alone, to accurately subtype a patient’s neuroanatomical pathology. Since functional connectivity scans on every patient would be expensive, it might be one day that the prudent approach would be to treat more than one network, either successively, after treating one network fails, or treating more than one area from the very beginning. With thetaburst potentially reducing treatment time to a few minutes, such an approach might yield the best outcomes with the fewest resources and lowest morbidity. All this is theoretical, of course, and should be tested in controlled studies.
Resting-State Connectivity Biomarkers Define Subtypes of Depression Study
Further evidence that functional connectivity might correlate with clinical symptoms and translate into defining depressive subtypes that helps clinicians individualized treatment
An important article called, “Resting-State Connectivity Biomarkers Define Subtypes of Depression” was published in Nature Medicine this year by Conor Liston and others. Based on patient symptoms and connectivity patterns, patients with depression were divided into four depression subtypes. The four subtypes or biotypes, were divided according to functional connectivity and symptoms, particularly the symptoms of anxiety, insomnia, psychomotor retardation, and anhedonia. Remarkably, Biotype 1 patients had an 80% response rate to TMS of the DMPFC (salience network involving cognitive control). Biotype 2 patients had only a 20% response rate to TMS. Biotype 2 has abnormalities in the OFC (orbitofrontal cortex), the non-reward network. Biotype 3 patients had a 60% response rate to TMS of DMPFC. Biotype 4 patients had only a 30% response rate to TMS, and tended to have more anxiety. Unfortunately, clinical features (insomnia, anhedonia and psychomotor retardation by HAMD) were only modestly (62.6%) predictive of treatment responsiveness, limiting, at this point, our ability to correlate symptoms with functional networks.
Towards a neuroanatomical formulation of depression
Today we are one step closer to a neuroanatomical formulation of depression based on functional connectivity, and to some extent, on clinical symptoms. With double-blind controlled studies of rTMS over the left DLPFC giving remission rates of 33%, and treatment of the right orbitofrontal region yielding remission rates of 25%, the combination would give us remission rates well over 50% with rTMS. In private practice, we often combine TMS with other treatment modalities, and using a targeted approach with TMS can give our patients a better chance for remission.