Brain Imaging in Psychiatry: MRI, CT, PET, SPECT, EEG, and Beyond

Psychiatry has lived for most of its history without being able to look directly at the organ it treats. The brain was, and largely still is, a black box — its biology inferred from behavior rather than observed in vivo. Brain imaging is the technology that began to crack that box open. Some modalities show the brain's anatomy as a still photograph. Others show its activity as a moving record. Others reveal its electrical chatter. Each was developed for a different purpose, became useful in psychiatry only later, and continues to find new applications. This chapter walks through the six major imaging technologies that touch psychiatric practice — CT, MRI, PET, SPECT, EEG, and functional imaging — describing how each works, when it was developed, what it can do for psychiatry today, and which ones look most promising for the next two decades.

Two Big Categories: Structural vs. Functional

Before walking through the modalities one at a time, it helps to keep two organizing axes in mind. The first axis is what the image shows. Structural imaging shows anatomy — gray matter, white matter, ventricles, masses, atrophy. Functional imaging shows activity — blood flow, metabolism, electrical or magnetic fields, receptor binding. The second axis is the underlying physical signal: ionizing radiation (CT, PET, SPECT), magnetic fields and radiofrequency waves (MRI, fMRI, MR spectroscopy), or electromagnetic activity directly emitted by neurons (EEG, MEG). Each modality combines one structural-vs-functional choice with one physical-signal choice, and the combination determines what the technique is good for, what it costs, and what its risks are.

The Six Modalities, One at a Time

Why Imaging Is Uniquely Useful for Psychiatry — Beyond the Obvious

The most obvious reason to image a psychiatric patient is to exclude a structural cause of the presentation: tumor, stroke, hydrocephalus, demyelinating disease, paraneoplastic syndrome, autoimmune encephalitis. That alone justifies a generous indication for MRI in atypical or first-episode presentations. But there are several less-obvious reasons imaging matters more in psychiatry than in some other specialties.

First, psychiatry treats conditions where the abnormality is distributed rather than focal. Heart failure, for instance, can be diagnosed and largely characterized by structure (echocardiogram) and function (ejection fraction). Schizophrenia and depression do not localize that way; they are network disorders. fMRI and MEG, which are good at characterizing network activity rather than focal pathology, are the modalities best matched to that biology.

Second, many psychiatric conditions involve subtle gray matter and white matter changes — frontotemporal atrophy, hippocampal volume loss, periventricular small-vessel disease, white-matter hyperintensities — that fall below the threshold for clinical certainty by examination alone. Quantitative MRI volumetry and diffusion tensor imaging extend the clinician's reach into territory that no examination can cover.

Third, psychiatric symptoms are often the earliest presentation of neurological disease. Depression and apathy in early Alzheimer disease, behavioral and personality changes in frontotemporal dementia, REM-sleep behavior disorder and visual hallucinations in dementia with Lewy bodies, and depression or psychosis in autoimmune encephalitis can all precede neurological diagnosis by months or years. Imaging is the bridge between psychiatric presentation and neurological etiology, and it is often the test that confirms or rules out a treatable underlying disease.

Fourth, imaging is increasingly required to deliver newer treatments safely. Anti-amyloid antibody therapy for Alzheimer disease requires baseline MRI and a defined surveillance schedule for ARIA. TMS planning increasingly relies on individual structural MRI for coil placement. Deep brain stimulation for treatment-resistant depression and OCD relies on structural and functional imaging for target selection. Imaging has thus moved, in selected indications, from diagnostic adjunct to therapeutic prerequisite.

Which Technologies Hold the Most Promise Going Forward?

Three modalities, in particular, are likely to shape psychiatric practice over the next two decades.

MRI, broadly conceived, is poised to remain the workhorse. Higher field strengths (7 T and beyond), faster sequences, and sophisticated machine-learning analysis pipelines are pushing structural and functional MRI from descriptive research tool toward individual-patient inference. Resting-state fMRI signatures of treatment response, functional connectivity-guided TMS targeting, and quantitative volumetry as a clinical biomarker are all moving from research into early clinical translation. Importantly, MRI uses no ionizing radiation, can be repeated longitudinally, and pairs well with multiple analytic frameworks.

Molecular PET imaging, especially for tau, synaptic density (SV2A ligands), neuroinflammation (TSPO), and selective receptor occupancy, will increasingly support both diagnosis and treatment. The clearest near-term win is in dementia: amyloid and tau PET have already moved from research into clinical care, and the arrival of disease-modifying therapy will probably entrench them. In psychiatric (rather than neurodegenerative) populations, PET ligands for synaptic density and microglial activation may yield the first true biological subgrouping of patients with depression, OCD, and schizophrenia.

EEG, MEG, and high-density electrophysiology have an underappreciated future. They are the only modalities with millisecond resolution; they are non-invasive; and they are increasingly portable. Mobile and dry-electrode EEG systems are bringing neural data outside the hospital, enabling long-term monitoring and treatment-response assessment in a way that scanner-bound modalities cannot. Combined with closed-loop neuromodulation devices, electrophysiology may eventually deliver real-time, individualized treatment in domains as varied as treatment-resistant depression, OCD, and epilepsy.

Two modalities are more cautiously optimistic. SPECT is unlikely to grow except in its established niches (DaT, seizure localization), as PET continues to take its market share. CT will remain essential in emergency triage but is unlikely to expand its psychiatric role.

The most important development, however, is probably not any single modality but the integration of imaging with other data. Multimodal pipelines that combine MRI, EEG, genetics, digital phenotyping, and clinical phenomenology — analyzed with modern machine-learning approaches — are starting to identify reproducible subgroups of patients within heterogeneous DSM categories. The promise is not that a single scan will replace clinical judgment, but that the next generation of psychiatric diagnoses will be more precise, more individualized, and more directly tied to mechanism than the categories we use today. Imaging will be one of several streams of evidence informing those diagnoses, just as the EKG is now one of several streams informing cardiology.

What This Means for the Practicing Psychiatrist

For the clinician working today, three habits track these developments. First, lower the threshold for ordering MRI in atypical, late-onset, or treatment-refractory presentations; the ratio of clinically meaningful findings to incidental findings is more favorable than many psychiatrists assume. Second, learn the appropriate use of EEG, especially when the differential includes seizure, encephalopathy, or autoimmune encephalitis — these are conditions where psychiatry holds the door to a neurological diagnosis that may be reversible. Third, treat published claims about "brain scans for psychiatric diagnosis" — particularly direct-to-consumer SPECT or qEEG services — with appropriate skepticism; the modalities are real, but their clinical validity for diagnosis in individual patients remains limited outside of specific, well-validated indications.

Imaging will not replace the psychiatric interview. It supplements it, the way an EKG supplements the cardiac examination — adding a complementary stream of evidence that the examination alone cannot provide. As the technology matures, the relationship will become more reciprocal: history and examination will guide imaging selection, and imaging findings will refine the differential diagnosis. The result, if the field gets it right, will be a psychiatry that is both more biologically grounded and more humanly responsive than the one its founders could have imagined.