Neuroimaging Methods — Differentiate structural and functional brain imaging methods and their applications

For most of the history of medicine, the living human brain was essentially a black box. Clinicians could observe behavior, assess cognitive function, and, in post-mortem examinations, correlate structural lesions with deficits—but they could not view the brain's architecture or activity in a living person. The advent of neuroimaging in the twentieth century transformed neuroscience and clinical practice by providing non-invasive windows into both the structure and the dynamic function of the brain. Understanding these methods is essential for behavioral health practitioners because diagnostic accuracy, treatment planning, and research interpretation all depend on knowing what each imaging modality can—and cannot—reveal.

The central question that neuroimaging addresses is deceptively simple: What does the brain look like, and what is it doing? Structural imaging answers the first question by depicting anatomy—ventricle size, cortical thickness, white-matter integrity, and the presence of tumors or lesions. Functional imaging answers the second by capturing metabolic activity, blood flow, or electrical signals that change when specific brain regions become active. As you prepare for the EPPP, the ability to differentiate these two broad categories, identify the specific modalities within each, and match them to clinical and research applications is a core competency in the Biological Bases of Behavior domain.

Neuroimaging modalities divide into two fundamental categories based on what they measure. Structural imaging techniques produce static or high-resolution anatomical pictures of the brain, revealing its physical composition—gray matter, white matter, cerebrospinal fluid, bone, and pathological tissue such as tumors or hemorrhages. Functional imaging techniques, by contrast, capture time-varying signals that reflect neural activity, blood flow, metabolism, or neurotransmitter dynamics. Within each category, individual modalities differ in spatial resolution (how small a structure or activation cluster they can detect), temporal resolution (how quickly they can capture changes), invasiveness, and sensitivity to particular tissue properties or physiological processes.

The diagram above highlights a key insight: the MRI platform is remarkably versatile. The same scanner that produces high-resolution structural images can, with different pulse sequences and analysis pipelines, generate functional maps (fMRI), white-matter tractography (DTI), and even measures of cortical perfusion. Other modalities, such as EEG and PET, rely on entirely different physical principles—electrical potential differences on the scalp and radioactive decay of injected tracers, respectively—and therefore provide unique but complementary information. Clinicians and researchers often combine modalities to compensate for the limitations of any single technique.

Structural Modalities

CT scanning works by rotating an X-ray source around the patient's head and measuring the attenuation of the beam as it passes through tissues of varying density. A computer reconstructs cross-sectional images from these attenuation profiles. Tissue density is quantified in Hounsfield units (HU), where water is 0 HU, bone is approximately +1000 HU, and air is −1000 HU. Because dense structures like bone and acute blood absorb more X-rays, CT is particularly effective for detecting skull fractures, acute hemorrhages, and calcified lesions, though its soft-tissue contrast is limited compared with MRI.

Structural MRI exploits the magnetic properties of hydrogen nuclei (protons), which are abundant in water and fat throughout the brain. When placed in a strong magnetic field (typically 1.5 or 3 Tesla), protons align with the field. A radiofrequency (RF) pulse tips them out of alignment; as they relax back, they emit detectable RF signals. The rate of relaxation differs between gray matter, white matter, and cerebrospinal fluid, producing excellent soft-tissue contrast. Two key time constants—T1 (longitudinal relaxation) and T2 (transverse relaxation)—are manipulated through different pulse sequences to emphasize particular tissue properties. T1-weighted images provide sharp anatomical detail (gray matter appears darker than white matter), while T2-weighted images highlight fluid-containing regions (cerebrospinal fluid appears bright), making them particularly useful for detecting edema and demyelinating lesions.

Functional Modalities

fMRI does not measure neural firing directly. Instead, it relies on the hemodynamic response—the increase in local blood flow and oxygenated hemoglobin that follows neural activity. Oxygenated hemoglobin (oxyhemoglobin) is diamagnetic, while deoxygenated hemoglobin (deoxyhemoglobin) is paramagnetic and distorts the local magnetic field. When neurons become active, oxygen consumption initially rises, but blood flow increases even more, creating a net surplus of oxyhemoglobin and a local increase in the MRI signal. This is the BOLD (Blood-Oxygen-Level-Dependent) contrast mechanism. The BOLD response peaks approximately 5–6 seconds after neural activation, which means fMRI's temporal resolution is inherently limited to the order of seconds rather than milliseconds.

PET (Positron Emission Tomography) introduces a biologically active molecule tagged with a positron-emitting radioisotope. The most common tracer is ¹⁸F-fluorodeoxyglucose (FDG), a glucose analog that accumulates in metabolically active cells. When the isotope decays, the emitted positron annihilates with a nearby electron, producing two gamma photons traveling in opposite directions. Coincidence detectors around the head reconstruct the spatial distribution of the tracer. PET is uniquely valuable for mapping neurotransmitter receptor density, studying metabolic abnormalities in disorders such as Alzheimer's disease, and detecting amyloid plaques using specialized tracers.

EEG (Electroencephalography) measures voltage fluctuations on the scalp produced by synchronous postsynaptic potentials in large populations of cortical neurons. Its temporal resolution is superb—on the order of milliseconds—making it ideal for studying neural oscillations, event-related potentials (ERPs), and sleep stages. However, because the electrical signal is smeared by the skull and scalp (volume conduction), EEG's spatial resolution is poor, on the order of centimeters. MEG (Magnetoencephalography) measures the tiny magnetic fields generated by the same neural currents. Because magnetic fields are less distorted by tissue, MEG offers slightly better spatial localization than EEG while preserving millisecond temporal resolution, though it requires expensive superconducting sensors.

No single neuroimaging method excels at both spatial and temporal resolution. This inverse relationship is one of the most important conceptual frameworks for evaluating imaging modalities. Techniques that precisely localize activity (high spatial resolution) tend to be slow, while those that capture rapid neural dynamics (high temporal resolution) sacrifice spatial precision. Understanding this trade-off helps clinicians and researchers choose the right tool for a given question.

This trade-off has practical consequences. A researcher studying the rapid time-course of language comprehension—where cortical processing unfolds within 100–600 milliseconds—would choose EEG or MEG despite their limited spatial resolution. A researcher investigating which specific brain regions activate during a memory retrieval task would opt for fMRI, accepting its slower temporal sampling. And a neurologist evaluating a patient for a suspected brain tumor needs structural MRI for its unmatched anatomical clarity, without concern for temporal dynamics at all. The EPPP frequently tests your ability to match clinical or research scenarios with the most appropriate imaging modality, making this trade-off framework essential.

In clinical and research settings, selecting the appropriate neuroimaging modality requires integrating knowledge of the patient's presentation (or research question), the physical characteristics of each technique, and practical constraints such as cost, availability, and patient safety. The following worked example walks through this decision process systematically.

Every neuroimaging modality carries a unique profile of strengths and limitations. The following table provides a comprehensive comparison that is highly relevant for both clinical reasoning and EPPP preparation. Note that no single technique is universally superior; the optimal choice always depends on the specific clinical or research question.

The neuroimaging techniques discussed so far represent the established toolkit of clinical and research neuroscience. However, the field continues to evolve rapidly, with new methods pushing the boundaries of what can be observed in the living brain. Awareness of these emerging developments is valuable for contextualizing current practice and anticipating the direction of the discipline.

One particularly significant development for behavioral health professionals is machine learning applied to neuroimaging data. Algorithms trained on large imaging datasets can classify brain scans as belonging to diagnostic categories (e.g., major depressive disorder vs. healthy controls) with increasing accuracy. While these tools are not yet standard in clinical practice, they represent a plausible future in which neuroimaging contributes directly to psychiatric diagnosis—a shift from the current reliance on behavioral assessment alone. Additionally, real-time fMRI neurofeedback is being investigated as a therapeutic intervention, where patients learn to modulate their own brain activity by watching their fMRI signal in real time. Early trials have explored this approach for chronic pain, PTSD, and substance use disorders.