Leanne Williams (standing) and psychiatry fellow Laura Hack review Moe’s brain scans while he looks on.

Leanne Williams (standing) and psychiatry fellow Laura Hack review Moe’s brain scans while he looks on.

LiPo Ching

Brain scans could help personalize treatment for people who are depressed or suicidal

By his late 20s, Moe had attained the young adult dream. A technology job paid for his studio apartment just blocks from the beach in Santa Barbara, California. Leisure time was crowded with close friends and hobbies, such as playing the guitar. He had even earned his pilot's license. "There was nothing I could have complained about," he says.

Yet Moe soon began a slide he couldn't control. Insomnia struck, along with panic attacks. As the mild depression he'd experienced since childhood deepened, Moe's life collapsed. He lost his job, abandoned his interests, and withdrew from his friends. "I lost the emotions that made me feel human," Moe says. (He asked that this story not use his full name.)

Although many people with depression respond well to treatment, Moe wasn't one of them. Now 37, he has tried antidepressant drugs and cycled through years of therapy. Moe has never attempted suicide, but he falls into a high-risk group: Though most people with depression don't die by suicide, about 30% of those who don't respond to multiple antidepressant drugs or therapy make at least one attempt. Moe was desperate for relief and fearful for his future. So when he heard about a clinical trial testing a new approach to treating depression at Stanford University in Palo Alto, California, near his home, he signed up.

People like Moe present a conundrum to doctors but an opportunity for researchers: a group whose health could be transformed by precision psychiatry. Depression is often treated as a single disease, but many researchers agree that it is actually multiple, distinct ailments. Some of those conditions may heighten suicide risk more than others. How many depression subtypes exist—and how they differ—is hotly debated. One way researchers are trying to settle the question is by peering into the brain. They're studying the neural circuits that light up during specific tasks and then correlating those patterns of activation with symptoms.

Those efforts are part of a broader campaign to explore the brain biology of mental illness, including depression, bipolar disorder, and active suicidality. The goal is not just to find biological markers of risk, but to tailor care accordingly—sometimes by reaching beyond psychiatry's usual armamentarium—and improve the prognosis of Moe and others like him.

The study Moe is part of, Research on Anxiety and Depression-Anhedonia Treatment (RAD-AT), sits on the leading edge of such efforts. It enrolls volunteers with a subtype of depression who are at higher risk of suicide than other types, and the study is among the first to offer treatments based partly on brain circuitry.

Whether such a targeted approach to depression can prevent suicide isn't clear. But researchers hope that linking symptoms to brain biology could help people who have languished despite treatment. Moe also longs for something else: insight into his own condition. "Therapists always [say], ‘Tell me the reason you feel this way,’" he says. "I want an answer."

The RAD-AT study is led by Leanne Williams, a Stanford clinical neuroscientist who has spent more than 20 years probing how depression manifests in the brain. She has orchestrated international collaborations to collect thousands of brain scans from depressed people. Like many in her field, Williams is driven to prevent depression's worst outcome: She lost a patient early in her career and, more recently, a loved one.

Thanks to her own and others' data, Williams believes at least six subtypes of depression exist. Each is generated by abnormal activity in a distinct set of brain circuits that regulate mood and cognition. One subtype affects a circuit called the default mode network, a constellation of brain regions that generates aimless mental chatter when the brain is "in idle" and can lead to unrelenting negative thoughts. Another type dampens reward networks, robbing a person of the ability to feel pleasure, a depression symptom called anhedonia. Those two subtypes, along with a third called cognitive control—which orchestrates attention, planning, and impulse control—often respond poorly to depression treatments, Williams says. She and others worry most about the anhedonia and cognitive control groups, partly because of their elevated suicide risk.

Scientists have already found several brain features that align with suicide risk. The best studied comes from neuroscientist John Mann of Columbia University. In the early 1980s, he examined the brains of people who had died by suicide, donated by their families. The organs had markedly lower levels of the neurotransmitter serotonin than those of depressed people who had died in other ways.

More recent work by him and his colleague Maria Oquendo, a psychiatrist at the University of Pennsylvania, suggests low serotonin levels may be more common in depressed people who attempt or die by suicide after struggling with persistent suicidal thoughts. That hypothesis is based on studies in which the pair used positron emission tomography (PET) imaging, which uses radioactive labels to track neurotransmitters, to capture serotonin levels in the brain. The team's data, which included a 2016 study of 100 depressed and suicidal people, suggest the low-serotonin group is biologically distinct from people who experience "spiky" bursts of suicidal thoughts during acute stress, such as a financial catastrophe or a breakup.

A June study in the Proceedings of the National Academy of Sciences, led by neuroscientist Irina Esterlis at Yale University, marked another step forward. It focused on post-traumatic stress disorder (PTSD), which can also raise the risk of suicide. Esterlis's group also used PET imaging. Among people with PTSD, those who had experienced suicidal thoughts had 30% more receptors for the signaling molecule glutamate, suggesting they were making less glutamate and the brain was struggling to compensate. The work pointed to a biomarker of suicide risk in people with PTSD. Her finding also engendered hope that ketamine, a drug that targets glutamate, might help people in that group. Recently approved as a rapid-acting antidepressant, ketamine is now being tested to see whether it can reduce suicidality.

Categorizing patients into neat buckets on the basis of neurotransmitter levels and other brain scan features is challenging. No matter what any scan measures and what struggles a patient faces, a scan is a snapshot in time. It can't reliably capture symptoms that wax and wane. "How does one pull up a brain readout of an emotional fluctuation like a surge of elation or anxiety or suicidal thoughts?" asks Helen Mayberg, a neurologist at Mount Sinai Hospital in New York City.

Depression subtypes can be parsed in many ways, and scientists don't agree on the best approach. Some researchers sort people based on how they respond to treatment, others according to symptoms. Volunteers are typically asked to engage in mental tasks, but those tasks can vary. Other researchers use biological markers as diverse as genetics, hormones, and gut bacteria to distinguish dozens of depression categories. In a 2016 study, on which Mayberg was an author, researchers analyzing 1000 functional magnetic resonance imaging (fMRI) scans of depressed people found four depression subtypes, not the six Williams has identified. Unlike PET scans, fMRIs measure brain activity by detecting changes in blood oxygen levels. But a separate group could not replicate the finding, Mayberg notes.

Because depression is so varied and complex, nailing down definitive categories could take many thousands of brain scans, says Elizabeth Ballard, a clinical psychologist at the National Institute of Mental Health in Bethesda, Maryland. But, "Everybody acknowledges that is what's needed," she says.

STANFORD UNIVERSITY

If the goal is to prevent suicides, some researchers also question whether probing depression is a good starting point, because most of those patients aren't at risk. Some scientists even argue that the biology of suicidal behavior is so different from that of depression—and can include symptoms of anxiety, agitation, and impulsivity—that "suicidal behavior" should be a stand-alone diagnosis. "We can't simply rely on the treatment of depression" to prevent suicide, Oquendo says, because suicidal behavior has different biological roots.

Williams agrees that suicidality and depression don't neatly align, but she rejects the notion that they must be studied separately. In some people, the two are undoubtedly intertwined, she says. Her partner, an emergency room doctor, had long struggled with depression but feared that seeking treatment could mar his professional reputation. Four years ago, he killed himself. The loss steeled her resolve to crack the biology of depression and improve its treatment. "Regardless of the labels," Williams says, "we need to look at where the crisis is."

For Williams, characterizing depression begins with the fMRI scanner, where her study volunteers lie, heads immobilized, while performing a battery of mental tasks. Each task exercises a different assembly of circuits that correspond to the six depression subtypes Williams has hypothesized are key to guiding treatment. The subtypes marked by repetitious negative thoughts and anhedonia are rooted in the default mode and reward circuits, whereas others involve circuits that respond to threat or help the brain maintain focus. When those circuits are dysfunctional, people may interpret events in a more negative light or feel trapped inside a mental "fog," Williams says.

The scan detects neuronal activity by measuring changes in blood oxygen levels, revealing how different regions of the brain fire and coordinate brainwide neuronal chatter. To account for individual variation in brain structure and activity, researchers must digitally "strip" the brain from the skull and align it to a standard model, says neuroimaging research engineer Brooke Staveland, who works with Williams at Stanford.

Sophisticated computer algorithms extract relevant patterns from the fMRI results and compare them with the healthy baseline. The result is a six-item chart that scores activity in each circuit, helping the researchers flag abnormalities.

Although Williams and her team have scanned patients with depression for years, the 8-week RAD-AT study goes a step further. It examines how 160 people with the anhedonia depression subtype, who often aren't helped by antidepressant drugs, respond to two treatments: transcranial magnetic stimulation, a noninvasive therapy that uses magnetic fields to stimulate nerve cells and is approved for treatment-resistant depression, and pramipexole, a drug for Parkinson's disease. Pramipexole mimics dopamine, the signaling molecule for the reward circuit that seems sluggish in those patients. In other studies, Williams is targeting additional subtypes, such as the tough-to-treat variants involving the default mode and cognitive control networks.

To qualify for RAD-AT, volunteers must score higher than normal on a standard questionnaire of anhedonia. Those who do are offered one of the two treatments. (Because pramipexole can increase impulsivity, actively suicidal people are excluded from the study for safety reasons.) Participants get their brains scanned beforehand to gauge activity in the reward and other depression-related circuits. After 8 weeks, they'll get scanned again to see whether treatment altered the circuits' activity and whether that change is associated with a change in symptoms.

Moe's brain scans are among those now being analyzed and considered alongside his clinical history. So far, the two appear to match up: He maxed out at an anhedonia score of 50, the highest possible, and had abnormally low activity on a task that activates the reward circuit: looking at photos of happy faces. If Moe's inability to experience pleasure is driven by too little dopamine, pramipexole could help, Williams says. In July, Moe agreed to take it.

What ultimately matters to Williams is not the number of depression or suicide subtypes, but how that knowledge helps patients. One barrier to widespread application is the time and expense of brain scanning. Williams is working to shorten the time to analyze a scan from a few hours to 5 minutes, and she and colleagues are weighing whether more easily tracked measures, such as heart rate, can serve as proxies for certain neuroimaging data. If so, the researchers hope to create wearable devices to help monitor depression, anxiety, and suicidal behaviors in real time. But first Williams needs more data, from her own lab and others, to determine whether differences in brain biology can translate to better treatment decisions.

RAD-AT is slated to end next year. Meanwhile, Williams and colleagues are running other neuroimaging studies, including one of 250 young people with depression that will explore how suicidal thoughts and prior attempts manifest in the brain. One young woman, who was actively suicidal, had refused medication because it failed her in the past. Brain scans suggested an abnormality in one of the three brain circuit types that don't respond well to antidepressants. She was then offered—and accepted—transcranial magnetic stimulation because previous studies showed its ability to correct abnormal activity in that network.

Her symptom scores and level of suicidality dropped into the healthy range. "I remember when she said to her mom, ‘I feel like myself,’" Williams says. Recasting depression as a disease of misfiring circuits can be a huge relief for people, she believes. Depression's stigma stopped her partner from seeking treatment. By lifting it, she hopes to make it easier for others to get help.

Moe is providing another, still provisional, data point. After 2 weeks on pramipexole, he felt better than he had in years. While driving to Stanford for an appointment, Moe switched on his car radio and heard the indie rock he'd loved in college. To his amazement, the music moved him. "I teared up for the first time in a long time, not because I was sad but because I was connecting with something again," he says.

Moe knows it's too early to tell whether what he describes as his reawakening will endure. If he continues to benefit, he'll keep taking pramipexole under a psychiatrist's supervision. But right now, he feels hope. "It's so weird," he says, "that you can take a medication and then wake up and say, ‘I think there's a future now.’"

For help, call 1-800-273-8255 for the National Suicide Prevention Lifeline, or visit https://www.speakingofsuicide.com/resources.