Neuroimaging

Neuroimaging is the set of techniques that let us see the living brain—its structure, its chemistry, and its activity—without opening the skull.

16 frameworks, from 1918 to today · Neuroscience · Natural Sciences

still a working view ended · each leads to its entry below

Overview

Neuroimaging is the set of techniques that let us see the living brain—its structure, its chemistry, and its activity—without opening the skull. What is at stake is direct: for most of history, knowledge of the human brain came from dissection of the dead and from the misfortune of the injured. Neuroimaging turned the working brain into an observable object. It made it possible to ask where in the brain a mental operation happens, how regions connect, and how activity changes with disease, development, and learning. The history below follows the instruments and the ideas they carried, from X-ray shadows of the skull to the maps of brain networks made today.

Shadows of the Skull: Contrast Agents and the First Living Images

The first step was not seeing the brain but seeing through it. Wilhelm Röntgen's discovery of X-rays in 1895 produced the first images of the living human interior, but the skull is dense and the brain soft, so plain radiographs showed bone and little else. The solution was to introduce something that absorbs X-rays differently from tissue.

In 1918 the American neurosurgeon Walter Dandy injected air into the ventricles of the brain—the fluid-filled chambers—and X-rayed the head1. Air is far less dense than tissue, so the ventricles appeared as dark outlines. This was pneumoencephalography (air injected into the spinal fluid spaces) and ventriculography1. It was painful and risky, but for decades it was the only way to see the brain's internal shape in a living patient, and it could reveal a tumor by displacing the ventricles.

A parallel line came from angiography. In 1927 the Portuguese neurologist Egas Moniz injected a contrast agent into the carotid artery and X-rayed the head, making the brain's blood vessels visible. Cerebral angiography showed aneurysms, malformations, and the displacement of vessels by masses. Both techniques were invasive and uncomfortable, but they established a principle that would recur: make the invisible visible by introducing a signal the body does not produce on its own.

Listening to the Brain's Electricity

A different tradition asked not what the brain looks like but what it does. In 1924 the German psychiatrist Hans Berger recorded electrical potentials from the human scalp and found a rhythmic oscillation—the alpha rhythm—that changed with eyes open and closed. He published the electroencephalogram (EEG) in 19292. EEG measures the summed electrical activity of large populations of neurons, and it gave clinicians a way to detect epilepsy and sleep stages and gave researchers a window on brain dynamics, though with poor spatial resolution because the skull blurs the signal.

Magnetoencephalography (MEG) came much later, in the late 1960s, when physicists led by David Cohen showed that the tiny magnetic fields generated by neural currents could be detected with sensitive superconducting sensors. MEG offered better spatial localization than EEG and, like it, millisecond timing. Both remain the tools of choice when the question is when, not just where1.

Mapping Metabolism: Radiotracers and Autoradiography

The next conceptual leap was to image not anatomy or electricity but chemistry—which parts of the brain are using energy. In the 1950s and 1960s, researchers developed radioactive tracers that cells take up in proportion to their metabolic activity. A landmark was the 2-deoxyglucose method, developed by Louis Sokoloff and colleagues in the 1970s: a glucose analog labeled with a radioisotope is taken up by active neurons but not fully metabolized, so it accumulates where energy demand is high. In animal studies, autoradiography—placing brain slices against film—produced detailed maps of activity.

This work established the metabolic logic that functional imaging would later exploit: active brain regions consume more glucose and oxygen and receive more blood. It also introduced the idea of cognitive subtraction, which would become central: if you image the brain during task A and during task A plus an added mental operation, the difference isolates the added operation.

Computed Tomography: Reconstructing a Slice

The decisive break came from mathematics and computing. In 1972 Godfrey Hounsfield, working at EMI, and independently Allan Cormack, showed that a series of X-ray measurements taken from many angles around the head could be combined by computer into a cross-sectional image—a slice through the body. This was computed tomography (CT)1. Hounsfield's first clinical scans of the brain revealed tumors and hemorrhages with a clarity that plain X-rays never approached, and CT quickly displaced pneumoencephalography. The key idea—reconstructing an image from projections—was general, and it opened the door to every subsequent tomographic method.

Positron Emission Tomography: Imaging Function in Humans

CT showed structure. The question of function in the living human brain was answered by positron emission tomography (PET). PET detects pairs of gamma rays emitted when a positron—a positively charged electron—annihilates with an electron in the tissue3. By labeling a molecule with a positron-emitting isotope, researchers could trace its distribution1. In the late 1970s and 1980s, groups led by Michel Ter-Pogossian, Michael Phelps, and others built PET systems, and researchers such as Marcus Raichle used labeled water and oxygen to measure regional blood flow and metabolism in the human brain during tasks.

The 1980s brought a wave of PET studies mapping language, memory, and attention. The method's spatial resolution was modest and its temporal resolution poor, and it required a cyclotron and radioactive tracers, but it was the first way to watch a healthy human brain perform a mental task. PET also enabled molecular imaging—receptor maps, dopamine studies—that remains one of its strengths1.

Magnetic Resonance Imaging: From Nuclei to Anatomy

Magnetic resonance imaging (MRI) grew from nuclear magnetic resonance, a physics technique for studying atomic nuclei in magnetic fields4. In 1973 Paul Lauterbur showed that magnetic field gradients could encode spatial position into the NMR signal, producing an image; Peter Mansfield developed the mathematical and practical methods that made fast, efficient imaging possible. The first human MRI images followed in the late 1970s5.

MRI's advantage over CT was contrast: different tissues—gray matter, white matter, cerebrospinal fluid—could be distinguished by tuning the pulse sequence, and there was no ionizing radiation4. By the 1990s MRI had become the standard structural imaging method in neuroscience and medicine, revealing anatomy with a precision that dissection could only approximate in the living.

Blood-Oxygen-Level-Dependent fMRI: Watching Activity

The most consequential development for cognitive neuroscience came from an observation about blood. In 1990 Seiji Ogawa and colleagues showed that deoxygenated hemoglobin distorts the magnetic signal, so changes in blood oxygenation produce image contrast678. This is the blood-oxygen-level-dependent (BOLD) signal68. Because active brain regions receive more blood than they consume oxygen, local oxygenation rises with activity7.

In 1992, several groups—including those of Kenneth Kwong and colleagues at Massachusetts General Hospital and Ogawa's collaborators—demonstrated that BOLD contrast could map brain activity during tasks in humans. Functional MRI (fMRI) was born1. It required no tracer injection, offered millimeter spatial resolution, and could be repeated in the same subject17. Within a decade fMRI had transformed cognitive neuroscience, producing thousands of studies mapping perception, language, memory, and emotion.

The method's limitations shaped its history too: the BOLD signal is slow (seconds), indirect (it measures blood, not neurons), and noisy7. These constraints drove the development of statistical methods7. Statistical parametric mapping (SPM), developed by Karl Friston and colleagues from the early 1990s, provided a framework for comparing images voxel by voxel—each small volume element—and correcting for the thousands of statistical tests involved. Cognitive subtraction, inherited from the metabolic tradition, remained the workhorse design, though it was increasingly criticized for assuming that adding a task adds only the operation of interest.

Diffusion MRI and Tractography: Mapping Connections

Structural MRI showed where tissue was; diffusion MRI showed how water moves through it. In the mid-1990s, researchers including Peter Basser and colleagues showed that the directionality of water diffusion reveals the orientation of white-matter fibers, because water diffuses more freely along axons than across them. This is diffusion tensor imaging9. From these measurements, tractography algorithms reconstruct the pathways connecting brain regions—a noninvasive approximation of the wiring diagram109. Tractography became central to studies of development, aging, and disconnection in disease, though its reconstructions are indirect and debated.

Near-Infrared Spectroscopy: A Portable Window

A quieter line of development used light. Near-infrared spectroscopy (NIRS) shines near-infrared light through the scalp and measures how much is absorbed by oxygenated and deoxygenated hemoglobin, exploiting the fact that these molecules absorb light differently111. Because near-infrared light penetrates bone and tissue a few centimeters, NIRS can monitor cortical blood oxygenation with portable, inexpensive equipment. It lacks the spatial resolution of fMRI and reaches mainly the cortex, but it brought functional imaging to infants, bedside patients, and settings where a scanner cannot go.

Cellular Imaging and Optogenetics: Watching Neurons Themselves

Meanwhile, a different scale of imaging was advancing. Calcium imaging uses fluorescent indicators that light up when neurons fire, letting researchers watch hundreds of individual cells in awake animals. Optogenetics, developed in the mid-2000s by Karl Deisseroth and colleagues, made neurons controllable with light by inserting light-sensitive proteins—channelrhodopsin and others—into their membranes. Combined with imaging, it allowed causal tests: activate a specific cell population and observe the effect on activity and behavior. Two-photon microscopy, developed in 1990 by Winfried Denk, James Strickler, and Watt Webb, made deep-tissue imaging of individual neurons possible. This cellular tradition is distinct from human neuroimaging but increasingly informs it, as the BOLD signal is interpreted against what cellular imaging reveals about the underlying neural activity.

The Present: Networks, Multimodal Imaging, and Open Questions

Today neuroimaging is plural. Structural MRI and CT remain the clinical backbone. fMRI dominates human cognitive neuroscience, increasingly analyzed in terms of brain networks—correlated activity across distributed regions—rather than isolated spots. Diffusion imaging maps connections; PET provides molecular specificity; EEG and MEG supply timing; NIRS extends functional imaging beyond the scanner. Multimodal studies combine these to compensate for each method's limits.

The open questions are the ones the history has produced. What does the BOLD signal actually reflect about neural activity? How reliable are tractography reconstructions10? Can imaging support causal claims about brain-behavior relationships, or only correlations? And the oldest question remains: whether measuring blood, electricity, and chemistry in living tissue can explain how the brain makes a mind.

References

  1. Neuroimaging (History, Magnetoencephalography (MEG) and Electroencephalography (EEG), Positron emission tomography, Functional magnetic resonance imaging, Diffuse optical imaging), Wikipedia
  2. Default mode network (History), Wikipedia
  3. Positron emission tomography, Wikipedia
  4. Magnetic resonance imaging, Wikipedia
  5. Magnetic resonance imaging of the brain (History), Wikipedia
  6. Seiji Ogawa, Wikipedia
  7. Functional magnetic resonance imaging (introduction, History), Wikipedia
  8. Blood-oxygenation-level–dependent imaging (History, introduction), Wikipedia
  9. Diffusion-weighted magnetic resonance imaging (introduction, Diffusion tensor imaging), Wikipedia
  10. Tractography, Wikipedia
  11. Functional near-infrared spectroscopy, Wikipedia

The frameworks of Neuroimaging

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1918-1975
  2. 1927-1985

    Cerebral Angiography

    Where do the brain's blood vessels run?

    Fly to its tower in Nousfarer

  3. 1929-Present

    Electroencephalography

    What electrical rhythms does the brain produce?

    Fly to its tower in Nousfarer

  4. 1930-Present
  5. 1955-1985
  6. 1968-Present

    Magnetoencephalography

    Where do neural currents generate magnetic fields?

    Fly to its tower in Nousfarer

  7. 1972-Present

    CT scan

    What does a slice through the head show?

    Fly to its tower in Nousfarer

  8. 1975-Present

    Positron Emission Tomography

    Which brain regions are active during a task?

    Fly to its tower in Nousfarer

  9. 1977-Present

    Magnetic Resonance Imaging

    How can tissue contrast be imaged without radiation?

    Fly to its tower in Nousfarer

  10. 1990-Present
  11. 1990-Present

    Functional near-infrared spectroscopy

    How does blood oxygenation change with activity?

    Fly to its tower in Nousfarer

  12. 1991-Present

    Statistical Parametric Mapping

    Which voxels differ significantly between conditions?

    Fly to its tower in Nousfarer

  13. 1992-Present

    Blood-oxygenation-level–dependent imaging

    Where does blood oxygenation track mental activity?

    Fly to its tower in Nousfarer

  14. 1994-Present

    Diffusion MRI and Tractography

    How are white-matter fibers oriented and connected?

    Fly to its tower in Nousfarer

  15. 2000-Present

    Multimodal Neuroimaging

    How can different brain images be combined?

    Fly to its tower in Nousfarer

  16. 2005-Present

    Brain Network Analysis

    How do brain regions form connected networks?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Pneumoencephalography and Ventriculog…Cerebral AngiographyElectroencephalographyElectrophysiological and Magnetoencep…Autoradiography and Metabolic Tracer …MagnetoencephalographyCT scanPositron Emission TomographyMagnetic Resonance ImagingCalcium Imaging and OptogeneticsFunctional near-infrared spectroscopyStatistical Parametric MappingBlood-oxygenation-level–dependent ima…Diffusion MRI and TractographyMultimodal NeuroimagingBrain Network Analysis

Beyond the field

Where Neuroimaging's ideas came from in other fields, and what they led to there: the strongest 12 of 76 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Neuroimaging is a planet in Nousfarer

Its city has a tower for each of these 16 frameworks. Land, walk in, and climb one concept at a time.

Land on Neuroimaging

Related fields

Fields its ideas are linked to

More in Neuroscience

See where Neuroimaging sits on the Map of Human Knowledge