Brain Imaging in Psychiatry: How fMRI, PET, and EEG Are Advancing Mental Health Neuroscience

For most of psychiatry's history, clinicians have diagnosed mental health conditions based entirely on observable behavior, self-reported symptoms, and clinical interviews. Unlike cardiology, which has the electrocardiogram, or oncology, which has biopsy and imaging, psychiatry has lacked objective biological markers for its diagnoses. Brain imaging is changing that — slowly, carefully, and with important caveats.

Neuroimaging technologies such as functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and electroencephalography (EEG) allow researchers and clinicians to observe the living brain in action. These tools have revealed that psychiatric conditions are not simply "chemical imbalances" or failures of willpower — they involve measurable differences in brain structure, function, connectivity, and neurochemistry.

However, the field is at a pivotal juncture. While research-level findings are robust and accelerating, the translation of brain imaging into everyday clinical psychiatric practice remains limited. Understanding what these technologies can and cannot tell us is essential for anyone interested in the neuroscience of mental health.

Each major brain imaging modality captures different aspects of brain activity, and understanding their mechanisms is critical to interpreting what they reveal about mental health.

Functional Magnetic Resonance Imaging (fMRI)

fMRI measures changes in blood oxygenation levels across the brain — a proxy for neural activity known as the blood-oxygen-level-dependent (BOLD) signal. When neurons in a brain region become active, they consume oxygen, triggering an increase in local blood flow. fMRI detects this hemodynamic response with spatial resolution of about 1-3 millimeters, producing detailed maps of which brain areas are engaged during tasks or at rest.

Resting-state fMRI is particularly important in psychiatry. Rather than requiring participants to perform tasks, it measures spontaneous fluctuations in brain activity while a person lies quietly in the scanner. This approach has revealed intrinsic brain networks — groups of regions that activate and deactivate in coordinated patterns — that are consistently disrupted across psychiatric conditions.

Positron Emission Tomography (PET)

PET imaging uses radioactive tracers — molecules labeled with short-lived isotopes — that are injected into the bloodstream. These tracers bind to specific targets in the brain, such as neurotransmitter receptors, transporters, or enzymes. As the isotope decays, it emits positrons that produce gamma rays detectable by the scanner.

PET is the only imaging technology that can directly measure neurochemical processes in the living brain. It can quantify dopamine receptor density, serotonin transporter availability, glucose metabolism, and neuroinflammation markers. This makes it uniquely valuable for understanding the molecular underpinnings of psychiatric conditions, though its use of radioactive materials limits how frequently it can be performed.

Electroencephalography (EEG)

EEG records electrical activity from the brain using electrodes placed on the scalp. It measures the summed postsynaptic potentials of large populations of neurons with millisecond-level temporal resolution — far superior to fMRI or PET, which operate on timescales of seconds to minutes. However, EEG has relatively poor spatial resolution and primarily detects activity from the cortical surface.

EEG is particularly useful for studying the timing and rhythmicity of brain activity. Specific frequency bands — delta (1-4 Hz), theta (4-8 Hz), alpha (8-13 Hz), beta (13-30 Hz), and gamma (30-100 Hz) — are associated with different cognitive states, and abnormalities in these rhythms have been linked to numerous psychiatric conditions.

Brain imaging research in psychiatry has consistently implicated several brain regions and large-scale neural networks in mental health conditions. While no single region "causes" a psychiatric disorder, patterns of dysfunction across interconnected systems help explain the complex symptom profiles clinicians observe.

The Prefrontal Cortex

The prefrontal cortex (PFC), particularly the dorsolateral and ventromedial subregions, is central to executive function, decision-making, emotional regulation, and working memory. Reduced prefrontal activity or volume has been documented in depression, ADHD, schizophrenia, and substance use disorders. The PFC acts as a "top-down" regulatory system — when its function is compromised, subcortical emotional and reward circuits operate with less oversight.

The Amygdala

The amygdala is a key hub for threat detection, fear conditioning, and emotional salience. Hyperactivity of the amygdala has been consistently observed in anxiety disorders, PTSD, and social anxiety disorder using fMRI. In depression, amygdala reactivity to negative stimuli is often heightened, while its connectivity with prefrontal regulatory regions is weakened.

The Default Mode Network

The default mode network (DMN) — comprising the medial prefrontal cortex, posterior cingulate cortex, and angular gyrus — is active during self-referential thought, rumination, and mind-wandering. Aberrant DMN activity is one of the most replicated findings in psychiatric neuroimaging. Hyperconnectivity within the DMN is associated with the rumination characteristic of major depressive disorder, while disrupted DMN connectivity has been observed in schizophrenia and autism spectrum disorder.

The Striatum and Reward Circuitry

The ventral striatum, particularly the nucleus accumbens, is a core component of the brain's reward system and receives dense dopaminergic input from the ventral tegmental area. PET studies have demonstrated altered dopamine signaling in this circuit across schizophrenia, addiction, ADHD, and depression. Reduced striatal activation during reward anticipation, measured by fMRI, is a hallmark finding in anhedonia — the inability to experience pleasure — which cuts across multiple diagnostic categories.

The Anterior Cingulate Cortex

The anterior cingulate cortex (ACC) plays a role in error monitoring, conflict detection, and the integration of cognitive and emotional processing. Structural and functional abnormalities of the ACC have been reported in depression, OCD, PTSD, and borderline personality disorder. The ACC appears to serve as a critical interface between the brain's cognitive control and emotional processing systems.

Decades of neuroimaging research have produced a substantial body of evidence linking specific patterns of brain activity and structure to psychiatric conditions. The following overview reflects well-established findings from meta-analyses and large-scale studies.

Major Depressive Disorder

fMRI studies consistently show reduced prefrontal cortical activity and increased amygdala reactivity in individuals with depression. Resting-state studies reveal hyperconnectivity within the default mode network, which correlates with the severity of rumination. PET research has identified reduced serotonin transporter binding in multiple brain regions, though findings vary depending on antidepressant history and illness duration. EEG studies have documented frontal alpha asymmetry — relatively greater left frontal alpha power (indicating reduced left frontal activity) — as a potential marker of depression vulnerability.

Schizophrenia Spectrum Disorders

PET studies provided some of the earliest biological evidence for schizophrenia, demonstrating elevated presynaptic dopamine synthesis capacity in the striatum. fMRI studies show reduced activation of the dorsolateral prefrontal cortex during working memory tasks — a finding so robust it is sometimes called "hypofrontality." Structural MRI reveals progressive gray matter loss, particularly in temporal and frontal regions. EEG studies consistently find reduced gamma-band oscillatory activity, which is thought to reflect impairments in cortical circuit function.

Anxiety Disorders and PTSD

Amygdala hyperactivation in response to threat-related stimuli is one of the most consistent fMRI findings across anxiety disorders. In PTSD specifically, the amygdala shows exaggerated responses to trauma-related cues, while the ventromedial prefrontal cortex — which normally downregulates amygdala activity — shows reduced engagement. This pattern is consistent with the clinical presentation of heightened fear responses and impaired fear extinction. EEG research in PTSD has identified abnormal patterns of frontal theta activity and disrupted sleep-related EEG architecture.

Attention-Deficit/Hyperactivity Disorder

Neuroimaging studies of ADHD have identified delayed cortical maturation, particularly in prefrontal regions, as well as reduced volume in the caudate nucleus and cerebellum. fMRI studies show hypoactivation in fronto-striatal and fronto-parietal networks during tasks requiring sustained attention and inhibitory control. PET and SPECT studies have demonstrated alterations in dopamine transporter density in the striatum, which is the target of stimulant medications.

Obsessive-Compulsive Disorder

OCD is associated with hyperactivity in the cortico-striato-thalamo-cortical (CSTC) circuit — a loop connecting the orbitofrontal cortex, striatum, and thalamus. PET studies show elevated glucose metabolism in the orbitofrontal cortex and caudate nucleus, and fMRI studies demonstrate excessive error-related signaling in the anterior cingulate cortex. Importantly, successful treatment — whether with SSRIs or cognitive-behavioral therapy — has been shown to normalize activity in these circuits, providing some of the strongest evidence that neuroimaging can track therapeutic change.

The field of psychiatric neuroimaging is evolving rapidly, with several research frontiers generating significant interest and cautious optimism.

The Transdiagnostic Approach and the Research Domain Criteria (RDoC)

Traditional psychiatric diagnosis, as outlined in the DSM-5-TR, groups patients by symptom clusters. But neuroimaging increasingly suggests that brain-based abnormalities cut across diagnostic boundaries. The Research Domain Criteria (RDoC) framework, introduced by the National Institute of Mental Health, encourages researchers to study dimensions of brain function — such as threat processing, reward sensitivity, or cognitive control — rather than categorical diagnoses. This approach has been validated by large neuroimaging studies showing that patterns like amygdala hyperactivity or default mode network dysfunction appear across depression, anxiety, PTSD, and other conditions.

Connectomics and Network Neuroscience

Rather than focusing on individual brain regions, modern neuroimaging increasingly examines the connectome — the comprehensive map of neural connections in the brain. Using techniques like diffusion tensor imaging (DTI) and resting-state fMRI, researchers can characterize the brain as a network and identify disrupted connectivity patterns associated with psychiatric symptoms. The Human Connectome Project and similar large-scale initiatives are generating datasets that allow unprecedented statistical power for these analyses.

Machine Learning and Predictive Neuroimaging

Researchers are applying machine learning algorithms to neuroimaging data in an effort to develop objective biomarkers for psychiatric diagnosis, treatment selection, and prognosis prediction. Some studies have achieved classification accuracy above 80% in distinguishing individuals with depression from healthy controls using fMRI data. However, these findings require rigorous external validation before clinical adoption. Guidelines such as the TRIPOD+AI statement emphasize the need for transparent reporting and independent replication of prediction models built on clinical and imaging data.

Neuroinflammation Imaging

PET tracers targeting the translocator protein (TSPO), a marker of microglial activation, have enabled researchers to visualize neuroinflammation in the living brain. Elevated neuroinflammation has been detected in individuals with major depression, schizophrenia, and bipolar disorder. This line of research supports the growing recognition that immune system dysregulation plays a role in psychiatric conditions, potentially opening new avenues for treatment.

Pharmacological Imaging

PET imaging is uniquely positioned to measure how psychiatric medications engage their brain targets. Studies measuring receptor occupancy — the proportion of receptors bound by a drug at a given dose — have directly informed dosing guidelines for antipsychotics and antidepressants. Research has established, for example, that optimal antipsychotic response typically occurs at 65-80% dopamine D2 receptor occupancy, while occupancy above 80% increases the risk of extrapyramidal side effects.

Despite remarkable research advances, it is essential to understand the current clinical reality of brain imaging in psychiatry.

What Brain Imaging Can Do

• Rule out neurological conditions: Structural MRI and CT scans are routinely used to rule out brain tumors, strokes, demyelinating diseases, and other neurological conditions that can present with psychiatric symptoms. This is the most established clinical application of brain imaging in psychiatric settings.
• Guide neurosurgical interventions: In treatment-resistant OCD and depression, neuroimaging is used to identify targets for deep brain stimulation (DBS) and guide electrode placement.
• Inform treatment response research: fMRI and PET findings are increasingly used in clinical trials to identify which patients are most likely to respond to specific treatments — a step toward personalized psychiatry.
• Support EEG-based clinical tools: Quantitative EEG (qEEG) has some FDA-cleared applications, including adjunctive tools for ADHD assessment and systems designed to help predict antidepressant response.

What Brain Imaging Cannot Do — Yet

• Diagnose psychiatric disorders: No brain imaging test can diagnose depression, anxiety, schizophrenia, or any other psychiatric condition. Group-level differences observed in research do not translate to reliable individual-level diagnosis. The overlap between "normal" and clinical brain scans is too great for any single scan to serve as a diagnostic test.
• Replace clinical evaluation: Psychiatric diagnosis remains fundamentally clinical, based on DSM-5-TR criteria applied through careful history-taking, behavioral observation, and clinical judgment. Brain imaging is a complement to — never a substitute for — thorough clinical assessment.
• Predict behavior with certainty: Despite media headlines, brain scans cannot reliably predict who will develop a mental illness, become violent, or respond to a specific medication at the individual level.

The gap between research findings and clinical utility is substantial. Many neuroimaging findings that are statistically significant at the group level have effect sizes too small to be clinically actionable for individual patients. Bridging this gap is one of the central challenges of translational neuroscience.

The public fascination with brain scans has produced widespread misunderstandings about what neuroimaging reveals. Correcting these misconceptions is important for both clinicians and the general public.

Misconception: Brain Scans Can Show If Someone Has a Mental Illness

This is perhaps the most pervasive and damaging misconception. Colorful fMRI images are visually compelling and convey a false sense of diagnostic certainty. In reality, the "hot spots" on an fMRI image represent statistical maps — averaged data from groups of people — not direct photographs of pathology. Individual brain scans from people with and without psychiatric conditions overlap substantially. Any clinic claiming to diagnose psychiatric conditions through brain scans is operating far beyond the evidence base.

Misconception: Mental Illness Is "Just" a Brain Disease

While neuroimaging demonstrates that psychiatric conditions have neural correlates, this does not mean they are reducible to brain abnormalities alone. Mental health conditions emerge from complex interactions among genetic vulnerability, neurodevelopment, psychological factors, social environment, and life experiences. Overemphasizing the brain disease model can inadvertently minimize the importance of psychotherapy, social support, and lifestyle factors in treatment and recovery.

Misconception: More Brain Activity Means Better Brain Function

The brain is not a muscle where more activation equals better performance. In many cases, excessive activation reflects inefficiency. For instance, the hyperactivation of the amygdala in anxiety disorders represents a maladaptive overresponse, not superior threat detection. Similarly, the hypermetabolism seen in the orbitofrontal cortex in OCD reflects a circuit stuck in overdrive, not enhanced function.

Misconception: We Only Use 10% of Our Brains

Neuroimaging has definitively debunked this myth. fMRI and PET studies show that virtually all brain regions are active at various points during the day. Even during sleep, the brain maintains widespread metabolic activity. There are no large, dormant regions waiting to be "unlocked."

Misconception: Brain Imaging Is a Routine Part of Psychiatric Care

Outside of ruling out neurological conditions, brain imaging is not a standard component of psychiatric evaluation. It remains primarily a research tool. While emerging clinical applications are being developed and tested, routine use of fMRI or PET for psychiatric diagnosis or treatment planning is not supported by current evidence-based guidelines.

The integration of neuroimaging into psychiatry raises important ethical questions that the field is actively grappling with.

Premature Commercialization

Several commercial entities market brain scans as diagnostic tools for psychiatric conditions, often using SPECT or qEEG technology. These services are typically expensive, not covered by insurance, and lack the rigorous validation required for clinical use. Leading professional organizations, including the American Psychiatric Association, have cautioned against the routine clinical use of brain imaging for psychiatric diagnosis outside of research settings.

Privacy and Neuroimaging Data

Brain imaging data is deeply personal — it can reveal information about cognitive function, emotional processing, and potentially even personality traits and predispositions. As neuroimaging datasets grow and machine learning techniques become more powerful, questions about data privacy, consent, and potential misuse become increasingly urgent. The WHO's guidance on ethics and governance of artificial intelligence for health emphasizes the need for robust data protection frameworks when AI tools are applied to sensitive health data, including neuroimaging.

The Replication Challenge

Psychiatric neuroimaging has faced significant challenges with replication. Many studies have used small sample sizes, employed inconsistent methods, and lacked rigorous correction for multiple comparisons. Large-scale collaborative efforts — such as the ENIGMA consortium, which pools neuroimaging data from thousands of participants worldwide — are addressing these limitations and producing more reliable findings. Recent mega-analyses from these consortia have confirmed some longstanding findings while failing to replicate others, leading to a healthier and more rigorous evidence base.

Toward Responsible Translation

The path from research finding to clinical tool requires multiple stages of validation: discovery, replication, prospective testing, clinical trials, and regulatory approval. Frameworks like the TRIPOD+AI guidelines provide standards for developing and reporting prediction models that use imaging and other clinical data, helping ensure that tools reaching clinical practice meet appropriate standards of accuracy and reliability. The FDA's guidance on clinical decision support software further clarifies the regulatory pathway for software tools — including those incorporating neuroimaging data — that assist clinical decision-making.

If you are experiencing persistent changes in mood, thinking, behavior, or daily functioning, the most important step is to seek a comprehensive evaluation from a qualified mental health professional — a psychiatrist, psychologist, or licensed clinical social worker. A thorough clinical assessment remains the gold standard for understanding mental health concerns.

You do not need a brain scan to seek help, receive an accurate diagnosis, or benefit from treatment. Evidence-based treatments — including psychotherapy, medication, and lifestyle interventions — are effective for a wide range of mental health conditions and do not require neuroimaging to guide their use.

In specific situations, a clinician may recommend brain imaging — typically structural MRI or CT — if there are signs that a neurological condition could be contributing to psychiatric symptoms. These signs include:

• New onset of psychiatric symptoms after age 50 with no prior history
• Sudden, dramatic changes in personality or cognition
• Neurological signs such as seizures, severe headaches, visual changes, or motor symptoms
• Symptoms that do not respond to standard psychiatric treatment as expected
• A history of significant head injury

If a provider recommends neuroimaging, it is appropriate to ask what specific clinical question the scan is intended to answer. If you encounter a provider or clinic marketing brain scans as a definitive diagnostic tool for psychiatric conditions, consider seeking a second opinion from an academic medical center or professional affiliated with a major psychiatric organization.

The neuroscience of mental health is advancing rapidly, and there is genuine reason for optimism about the future role of brain imaging in improving psychiatric care. In the meantime, the most effective path to better mental health begins with honest conversation, professional evaluation, and evidence-based treatment.