Schizophrenia Podcast

Neuroimaging in Schizophrenia: What Have We Learned?

John M. Kane, MD; Anil K. Malhotra, MD


November 09, 2022

This transcript has been edited for clarity.

John M. Kane, MD: I am Dr John Kane, professor of psychiatry and molecular medicine at the Donald and Barbara Zucker School of Medicine in Hempstead, New York. Welcome to Medscape's InDiscussion series on schizophrenia. Today we'll be discussing neuroimaging in schizophrenia. What have we learned?

First, let me introduce my guest, Anil Malhotra, who is co-director of the Institute of Behavioral Science at the Feinstein Institutes for Medical Research at Northwell Health and director of psychiatry research at the Zucker Hillside Hospital in New York. He has a long track record of research in schizophrenia, with over 400 peer-reviewed publications, and he currently holds three National Institutes of Health (NIH) grants focused on neuroimaging in schizophrenia. Anil, welcome to InDiscussion.

Anil K. Malhotra, MD: Thank you, John.

Kane: It's been quite a few years since the first neuroimaging studies in schizophrenia revealed things like enlarged ventricles. We've come a long way since that, with the introduction of MRI and a lot of new techniques for not just structural but functional imaging. Can you take us through the progress that we've made in recent years and what you are up to these days in terms of neuroimaging research?

Malhotra: As you mentioned, the early studies in schizophrenia focused on volumetric findings, such as enlarged ventricles and reduced hippocampal volume. But more recently, there's been a lot more attention paid to functional connectivity in schizophrenia and other psychiatric disorders. This is usually assessed with something called resting state MRI, where you can look at the activity of any number of different regions of the brain and see if there are correlations between different regions in the brain and the amount of activity. If you see correlations like that, we believe that suggests there's functional connectivity and/or circuitry between these regions of the brain. There is a lot of focus right now on trying to understand whether that functional circuitry or functional connectivity impacts either the signs and symptoms of schizophrenia or, in our work, treatment response in schizophrenia.

Kane: Can you tell the listeners what "resting state" actually means?

Malhotra: When we do an MRI scan, a patient will lie in the MRI scanner, which is basically a long tube, and be asked to rest, do nothing. We do not want the person to go to sleep. We want them to be awake, but hopefully thinking about very little, not engaging in a cognitive or behavioral task. So we get an examination of the brain while it's not performing some kind of cognitive or behavioral or emotional task. They're at rest, as opposed to a task-based MRI, where an individual is asked to do a cognitive or behavioral task, some kind of mathematical activity or something like that. In this case, the individual is strictly at rest. So we can look at the connectivity in the brain while the subject is resting.

Kane: I guess that implies that if someone is concentrating on a mathematical problem, that changes the connectivity that one would see.

Malhotra: Exactly. When we start to think about certain activities or do certain tasks, various regions of our brain are activated, and if you will, light up on an MRI scanner. We actually don't want that with resting state MRI. We're just trying to look at what the brain looks like when the patient is doing nothing, or again, at rest.

Kane: Can you tell if someone falls asleep?

Malhotra: That's actually a very big challenge. It's the great question. In reality, we cannot necessarily tell when someone falls asleep. Some scanners have cameras, and we can actually look at the patient while they're in the scanner and get very suspicious if their eyes are closed. We sometimes measure respiration rate, and if the respiration rate starts to get very slow and even and the patient's eyes are closed, we of course become suspicious that they're sleeping. We can speak to the person while they're in the scanner and ask them, are you awake? And hopefully they will respond yes. If they do not, then we are a little bit suspicious that they may be asleep, but it's a bit of a confounder in some cases.

Kane: Obviously you're studying patients with schizophrenia. Do you sometimes have trouble with the person being at rest in the scanner? Do they move? Do they get anxious?

Malhotra: Yes. Actually, not just patients with schizophrenia. Almost all individuals have some types of challenges with lying in a scanner and being told to do nothing. People fidget. People think about things. They may be worried about what's going on outside. A patient with schizophrenia, of course, may have other worries or be paranoid about what's going on. So the actual, at-rest state is certainly confounded a little bit by what could be going on in the brain. Having said that, when the original resting state studies came out, there was a bit of a concern about the reliability of resting state MRI. And one study looked at people 2 days apart. They were scanned on day one in the resting state and then scanned again a day or two following the first scan. And the scans were very, very reliable. Almost in every case, a trained expert could reveal who the person was from their first scan by their second scan. It's actually a very reliable index.

Kane: Tell us more about how your current research has evolved, and how did this start? What were the first observations in terms of resting state connectivity?

Malhotra: As I said, we start off with structural MRI. We then did some studies called diffusion tensor imaging, which looks at white matter integrity. And those studies have been helpful. But the first studies with resting states started to show different connectivity patterns in patients with schizophrenia compared with people without schizophrenia. In our case, we focused on something called striatal connectivity, the connections between the striatum in the brain and the frontal cortex. We were able to show that some differences in striatal connectivity impacted signs and symptoms of schizophrenia, but also, more importantly in our work, the treatment response to antipsychotic drugs in schizophrenia.

Kane: Why is the striatum so important?

Malhotra: The striatum is thought to be a very key region of the brain in schizophrenia for one major reason. The dopamine D2 receptor, which has been implicated in schizophrenia and in the treatment of schizophrenia, is highly located in the striatum of individuals. And so the striatum has had a lot of focus for the signs and symptoms of schizophrenia in pathophysiology of schizophrenia, but also for treatment response. Because as of today, all of our drugs work by binding to the dopamine D2 receptor. And again, the dopamine D2 receptor is very, very common in the striatum of individuals.

Kane: How did you go about studying treatment responses? Does that mean you did scans before and after the patients were treated? How did that work?

Malhotra: Exactly. As you're well aware, we have a program here in first episode schizophrenia, where patients who are in their first break are admitted to the hospital and hopefully they consent to participate in our research study. As part of our research study, we do an MRI scan in patients before they begin treatment with an antipsychotic drug. We do the MRI scan, which includes a number of different measures, including resting state MRI, and then they begin treatment with an antipsychotic drug. In our case, we do a 12-week treatment trial with either risperidone or aripiprazole. Patients go through the trial. We're seeing them every week during the trial. At the conclusion of the 12-week trial, we do another scan and then we can therefore compare the baseline scan before the treatment with the scan after 12 weeks of treatment.

Kane: You mentioned aripiprazole and risperidone. Is it important that you try to study the same drug whenever possible at a comparable dose? What would happen if patients got very, very different drugs?

Malhotra: Yes. We think it is important to study patients on the same drug, because if patients are switching drugs or at very different high or low doses, it may make it more difficult to find out what the actual effects of the drug are on the brain. So, we do try to standardize the treatment in these research studies as much as possible.

The other thing that's also very important, as you're aware, is that we have to make sure the patients are actually taking the antipsychotic drug. We prescribe the drug, but we have no way of knowing sometimes whether the patient is actually taking a drug. That can be a big confounder, too. We do blood levels of the antipsychotic drugs. So, we're pretty reassured by having a positive blood level that the patient is taking the drug. But you're still never precisely sure exactly how much of the drug they're taking, despite what we may hope the patient is taking.

Kane: Tell us more about what you found and in terms of measuring or predicting drug response.

Malhotra: In our first major study of striatal connectivity, we found that the amount of striatal connectivity actually correlated with the efficacy of the antipsychotic drug. From the baseline scan and then to the 12-week scan, we saw connectivity increasing between the striatum and the frontal cortex in key regions of the brain. And as we saw those patients who had the greatest increases in connectivity, those are the patients who had the greatest decreases in their psychotic symptoms. It was a very nice correlation between striatal connectivity and response to the antipsychotic drugs — risperidone and aripiprazole in this case.

In the second study, we looked at the amount of striatal connectivity at baseline and asked the question, does it predict who does well or not well on the antipsychotic drug? We found that striatal connectivity was a pretty good predictor of who was going to do well or not well on the antipsychotic drugs. Those patients with lower striatal connectivity at baseline were the patients more likely to respond to 12 weeks of treatment with the antipsychotic drugs.

Kane: So the more dysfunctional their connectivity was, the more likely they were to respond?

Malhotra: Exactly. And that correlates with the first paper, where we saw the amount of increase was actually correlated with the decrease in psychotic symptoms. Having a dysfunctional striatal connectivity system predicted that you might respond better to the dopamine D2 receptor antagonists that were used in our trial.

Kane: Does this mean that you could identify patients who were less likely to respond to a drug like aripiprazole or risperidone and may be a candidate for clozapine?

Malhotra: Yes. That's actually a big area of research for us now, because if you're unlikely to respond to the first-line agents, perhaps you should be treated with a drug like clozapine earlier, because obviously being on a drug that's not going to work has a lot of morbidity and perhaps mortality associated with it. The idea of going to clozapine earlier is an area we are actively investigating.

The other area, of course, is that there may be drugs down the road that don't act via the dopamine D2 receptor or via the striatum. Perhaps those patients also would be prioritized for treatment with those kinds of drugs. If you have a normal striatal connectivity system, that may not be what's driving your symptoms. Maybe you should be treated with a drug that does not act on striatal connectivity. We don't have those drugs as of yet, but as you're aware, there are many drugs in the pipeline that may act via non-D2 or nonstriatal mechanisms. Maybe those patients with the normal striatal connectivity systems may be prioritized for treatment with those agents.

Kane: You're referring to the M1/M4 agonist xanomeline and also the TAAR1 agonists. As you mentioned, there are some drugs in those categories under development. I guess there haven't been any resting state studies yet with those drugs, right?

Malhotra: Not that I'm aware of. We would very much like to do one. If those drugs are made available to us, that would be one of the very first studies that we'd embark on to test the specificity of the relationship between striatal connectivity and response to those non-D2 nonstriatal drugs. We're not quite there yet, but I very much look forward to the opportunity to do that kind of study.

Kane: What about electroconvulsive therapy (ECT)? What effect does that have?

Malhotra: ECT generates a generalized seizure, and so it has very large effects across the brain. Nothing is focused. For example, striatal connectivity tends usually to depress connectivity across the brain, at least in the short term. We're studying ECT in schizophrenia, looking at the effects of ECT on patients in whom clozapine has failed. A paper that you and I were involved in many years ago by Georgios Petrides found that of patients in whom clozapine failed and who had ECT after that, about 50% of them responded to treatment. Of course, that means 50% didn't respond to treatment. We're doing a resting state connectivity study right now, where we're looking at patients who are treated with ECT and trying to understand, is there a resting state signature of who is going to be more likely to respond to ECT or less likely to respond? That 50/50 response rate actually makes a very nice design for a study because we have pretty good power to hopefully detect that. We're about 3 or 4 years into that study. We don't have any data yet, but I'm very hopeful that we'll be able to predict response to ECT in schizophrenia using these resting state measures.

Kane: And as you mentioned earlier, in terms of the potential advantages of having a biomarker like resting state connectivity, what about predicting length of hospital stay or something like that?

Malhotra: In the paper that we did looking at striatal connectivity and treatment response, we also did look at length of hospital stay and days of hospitalization. The striatal connectedness did predict who is more likely to have a long hospital stay vs a short hospital stay. So that was quite nice. That was pretty consistent with the treatment response data. Of course, those patients who failed to respond to treatment also tended to be in the hospital longer, and so it did predict days of hospitalization.

Kane: Let's talk for a minute about another aspect of treatment nonresponse. That is the patients who have relapse despite taking antipsychotic medication. You mentioned earlier that one of the big challenges in doing this kind of research is being absolutely certain that the patient is taking the medicine. So, you've done blood levels to document that. But, another strategy is using long-acting injectable antipsychotics in the maintenance phase so that we know the person is actually getting the medicine. What applicability is there for the resting state connectivity marker in identifying those patients who might be vulnerable to relapse?

Malhotra: Jose Rubio in our group is doing just that study. He's looking at patients who have relapse while they're on long-acting injectable agents and is asking the question, does striatal connectivity have an impact or predict who's more likely to relapse following long-acting injectable medications? And again, yes, he finds a relationship between striatal connectivity and relapse while these patients are on guaranteed medication delivery. He's comparing a number of different groups. He's comparing patients who have relapse on medication vs patients who have relapse following medication discontinuation vs patients who don't have relapse while on long-acting injectables. He has an NIH grant now to study that question of whether striatal connectivity predicts those differences. And at least preliminarily, the answer is yes.

Kane: To change focus for a second, there's a lot of interest in trying to identify young people who might be in the so-called prodromal phase of schizophrenia — that is before they have a full-blown psychotic episode. We see some decline in school functioning and social interactions and perhaps even cognitive performance. What about neuroimaging? Can that play a role in helping to identify people at risk of developing a psychotic disorder like schizophrenia?

Malhotra: There has been quite a bit of research in that area. There was a consortium called NAPLS, the North American Prodrome Longitudinal Study, that was conducted several years ago that did neuroimaging in patients who were perhaps likely to convert to full-blown psychosis in the prodrome. And there were some signatures, mostly volumetric signatures of patients being more likely to convert to psychosis from the prodromal state with neuroimaging. Not enough really be a predictor of use in the clinic as yet, but certainly an intriguing finding.

Now there's a follow-up study called PRONIA, which you lead, looking again at the relationship between various neuroimaging and other biomarkers, trying to identify who is more likely to convert to illness from this prodromal state. That, of course, would have enormous utility if it could work out, because in the prodromal studies, most of the prodromal subjects do not ultimately convert to schizophrenia. The question is, should you be treating those patients earlier, even perhaps before the onset of illness? If you had a biomarker that lets you know who's really likely to convert to full-blown schizophrenia, I think you'd be more likely to treat those patients with antipsychotic or other treatment strategies. That kind of work is ongoing, and I think a lot of people are looking forward to the results.

Kane: There's obviously been a lot of discussion about the differential diagnosis of schizophrenia. We see someone with the first episode of psychosis, and I know at times there have been suggestions that we do neuroimaging more routinely, but that's not the general recommendation. Do you have any thoughts about that and also incidental findings that you've seen in some of your studies?

Malhotra: When I was trained, I remember Daniel Weinberger used to say that everybody who is in a first episode of psychosis should have a brain scan, and in those days we were talking more about CT scans than an MRI. The idea was that perhaps some of the patients who present with schizophrenia-like symptoms might have some kind of brain lesion that could be identified with CT. It turned out ultimately that this strategy was pretty low-yield. There were very, very low numbers of incidental findings in patients with first-episode psychosis. I don't know that anyone really carried out a full-blown study where they looked at tens of thousands of patients. For the most part, the number of incidental findings were very low.

The recommendation now is that patients with a first episode of psychosis are not routinely scanned with MRI or CT or anything. I will say, though, in our hands, we've had a few incidental findings in our first-episode research program, more than the predicted number that you would think. We've had a couple of incidental findings that have actually been actionable, and we've referred patients to neurosurgeons for some kind of surgical procedure. I can't say that it's completely resolved their symptoms, but the idea of doing scans at baseline is still probably not generally recommended at this point in time. Perhaps as we get better and better scanning methodologies and it gets easier and easier to do, that may be revisited and we may think about doing scans at baseline, but currently we don't do that.

Kane: We've talked about a number of different applications for neuroimaging and schizophrenia. Is there anything that we haven't mentioned or anything that you see as perhaps coming down the road in the future?

Malhotra: What I'm hopeful about is that the scanning technology gets easier and easier to do. You now have these mobile scanners that are available that can do a resting state scan in as short as 5 minutes. As that becomes easier and easier to do and hopefully cheaper and cheaper to do, the large-scale application of neuroimaging to schizophrenia and other psychiatric disorders may become more common because it is increasingly recognized that large sample sizes are really necessary for this kind of work. So, the boutique studies of 30 or 40 subjects that have been done may not be that relevant for much longer. Studies of hundreds, maybe thousands of subjects scanned, maybe at the same place and with the same scanner, will be more useful. As the methodology for scanning becomes easier and easier, I think we'll start to see that over the coming years.

I think it's an exciting time to be doing neuroimaging in schizophrenia. The idea of using these kinds of techniques, particularly resting state scanning, looking at connectivity between various regions of the brain and how that impacts the pathophysiology of the disorder and how it impacts treatment response in a disorder, I think, are areas of active investigation depending on the treatment — whether it be a D2 receptor antagonist, these new agents that are out there, or, as you mentioned, neuromodulatory techniques like ECT. And we did talk about transcranial magnetic stimulation (TMS) also, but can we use these techniques to understand what the best treatment strategy is for an individual suffering from schizophrenia? Can we predict what the best treatment strategy is and use that to alleviate the suffering of our patients with schizophrenia?

Kane: It's obviously a very exciting area, as you said. Thanks so much for joining us and thanks to the audience for participating. This is Dr John Kane for InDiscussion.



Diffusion Tensor Imaging Measures of White Matter Compared to Myelin Basic Protein Immunofluorescence in Tissue Cleared Intact Brains

Baseline Striatal Functional Connectivity as a Predictor of Response to Antipsychotic Drug Treatment

Dissecting Heterogeneity of Treatment Response of First Episode Schizophrenia

The M1/M4 Preferring Muscarinic Agonist Xanomeline Modulates Functional Connectivity and NMDAR Antagonist-Induced Changes in the Mouse Brain

Therapeutic Potential of TAAR1 Agonists in Schizophrenia: Evidence from Preclinical Models and Clinical Studies

ECT and Other Procedures for Schizophrenia

Electroconvulsive Therapy Augmentation in Clozapine-Resistant Schizophrenia: A Prospective, Randomized Study

Long-Acting Antipsychotic Therapies for Patients With Schizophrenia

Psychosis Relapse During Treatment With Long-Acting Injectable Antipsychotics in Individuals With Schizophrenia-Spectrum Disorders: An Individual Participant Data Meta-analysis

North American Prodrome Longitudinal Study

Association Between Residential Instability at Individual and Area Levels and Future Psychosis in Adolescents at Clinical High Risk From the North American Prodrome Longitudinal Study (NAPLS) Consortium

Clinical, Brain, and Multilevel Clustering in Early Psychosis and Affective Stages

Brain Disease and Psychiatric Illness: When Should a Psychiatrist Order a CAT Scan?

Transcranial Magnetic Stimulation for Positive Symptoms in Schizophrenia: A Systematic Review

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.