Neuroimaging, or brain scanning, refers to techniques that produce images of the brain’s structure or activity. These tools allow researchers and clinicians to observe the brain in action and identify abnormalities without surgery.
Neuroimaging is used in both clinical and research settings.
It helps us understand how different parts of the brain work, how disorders affect brain function, and how treatments influence brain activity.
Why Do We Use Neuroimaging in Psychology?
Neuroimaging is especially valuable in psychology and neuroscience for exploring how thoughts, emotions, and behaviors link to brain regions. It can:
- Show differences in brain structure and activity between individuals.
- Help diagnose brain injuries, tumors, or disorders like Alzheimer’s.
- Map brain function in real time (e.g., watching how the brain responds during memory tasks).
This non-invasive window into the brain allows scientists to test theories about cognition, behavior, and mental illness with unprecedented detail.
Types of Brain Imaging Techniques
Most neuroimaging techniques fall into two main categories:
- Structural imaging reveals the brain’s anatomy (e.g., MRI, CT).
- Functional imaging tracks brain activity in real time (e.g., fMRI, PET, EEG).
Below are some of the brain imaging techniques:
Structural Neuroimaging
These scans provide detailed images of brain anatomy and are often used to detect abnormalities like tumors, bleeding, or injury.
MRI (Magnetic Resonance Imaging)
- What it does: Uses strong magnetic fields and radio waves to create high-resolution images of brain tissue.
- Used for: Identifying brain tumors, multiple sclerosis, stroke damage, or brain malformations.
- What to expect: You’ll lie still inside a long tube. The scan is painless but noisy; you may be given earplugs. No radiation is involved. It typically takes 30–60 minutes.
CT Scan (Computed Tomography)
- What it does: Uses X-ray images to create cross-sectional views of the brain.
- Used for: Detecting bleeding, skull fractures, and brain swelling — especially in emergencies.
- What to expect: You’ll lie on a table that moves through a large ring-shaped scanner. The process is fast (5–10 minutes), and the machine is quiet. Some scans may require a contrast dye injection.
Functional Neuroimaging
These scans show brain activity and are useful in research and diagnosis of cognitive or neurological disorders.
fMRI (Functional Magnetic Resonance Imaging)
- What it does: Measures blood oxygen levels to map active areas of the brain during tasks or rest.
- Used for: Studying memory, emotion, attention, decision-making, and brain disorders like depression.
- What to expect: Similar setup to an MRI. You may be asked to complete simple tasks (like viewing images or tapping fingers) while the machine tracks brain responses.
PET Scan (Positron Emission Tomography)
- What it does: Uses a small amount of radioactive tracer to measure metabolic activity in brain tissue.
- Used for: Diagnosing Alzheimer’s disease, epilepsy, cancer, and some psychiatric conditions.
- What to expect: A radioactive tracer is injected into your bloodstream. After a short wait, you’ll lie on a table while the scanner captures images. The scan itself takes around 30 minutes and is painless.
Key Study: Localising Language (Petersen et al., 1988)
PET produced one of the field’s founding demonstrations: proof that a purely mental task could be mapped onto specific brain regions without surgery or lesions.
- Aim: To identify the separate brain regions supporting each stage of processing a single word, from seeing it to generating its meaning.
- Method: Healthy volunteers were scanned with PET while performing tasks of rising difficulty: viewing or hearing a word, reading it aloud, then generating a related word. Each step added one extra mental process to the last.
- Results: Different tasks lit up different, replicable regions: sensory areas for perceiving words, motor areas for saying them, and frontal regions for generating meaning.
- Conclusion: Complex mental tasks can be broken into components and mapped onto distinct brain regions in healthy, awake people (Petersen et al., 1988).
The study also planted an assumption later criticised: that a task cleanly isolates one process for a scanner to localise, when most regions actually support several processes at once.
EEG (Electroencephalography)
- What it does: Records electrical activity in the brain using sensors placed on the scalp.
- Used for: Monitoring brain waves, diagnosing epilepsy, sleep disorders, or attention-related conditions like ADHD.
- What to expect: Electrodes are gently placed on your scalp using a cap or adhesive. The test is quiet, non-invasive, and typically lasts 20–60 minutes.
MEG (Magnetoencephalography)
- What it does: Measures the magnetic fields produced by neuronal activity to pinpoint where and when brain processes occur.
- Used for: Mapping sensory functions, diagnosing epilepsy, and guiding brain surgery.
- What to expect: You’ll sit or lie with your head positioned near a helmet-like device. It’s silent, non-invasive, and often used alongside MRI for detailed brain mapping.

How Brain-Imaging Techniques Differ
Beyond structure versus function, two distinctions decide which scanner suits a question. One is precision in space and time. The other is whether neurons are measured directly or indirectly.
Spatial vs. Temporal Resolution
Spatial resolution is how finely a technique can localise activity in the brain, usually measured in millimetres. Temporal resolution is how finely it can localise activity in time, measured in seconds or milliseconds.
No technique excels at both: fMRI and PET pinpoint activity to a few millimetres. That is the trade-off. But they resolve time only to the nearest second.
EEG and MEG do the opposite: they track the brain’s electrical and magnetic signals directly, so they resolve timing to the millisecond. But they localise the source only roughly.
The question being asked decides the tool. Researchers reach for fMRI or PET to find out where a mental process happens. They reach for EEG or MEG to find out when it happens, and in what order.
Some studies combine both, pairing fMRI’s spatial precision with EEG’s timing on the same participants.
Direct vs. Indirect Measures: the BOLD Signal
A second, deeper distinction cuts across the functional techniques. Direct measures, such as EEG and MEG, pick up the electrical or magnetic signal of neurons firing in something close to real time.
Indirect (haemodynamic) measures, such as fMRI, PET and fNIRS, do not detect firing at all. They detect something else instead.
These measures detect the brain’s blood response a few seconds after neurons fire. The signal fMRI relies on is the BOLD signal: deoxygenated blood distorts the scanner’s field; oxygenated blood does not. A harder-working region floods with oxygenated blood, brightening the BOLD signal (Ogawa et al., 1990).
This matters for how scan results should be read. The BOLD signal is a delayed, blurred echo of neural activity, several seconds behind the neurons that triggered it. It is a proxy, not a photograph of a thought.
Together, these two distinctions (resolution and directness) explain why no single brain-imaging technique does everything, and why researchers often combine methods to answer one question fully.
History of Early Neuroimaging Techniques
Before the invention of MRI or fMRI, scientists used much cruder methods to study the brain.
While some early approaches are now outdated or discredited, they helped shape the evolution of modern neuroimaging techniques.
Phrenology (1800s)
One of the earliest (and most controversial) attempts to understand the brain was phrenology. This pseudoscientific theory claimed that bumps and indentations on the skull could reveal a person’s character traits and mental abilities.
Though inaccurate, phrenology introduced the idea of localization of brain function—the notion that different brain areas serve different purposes—a principle still central in neuroscience today.
Human Circulation Balance
Italian physiologist Angelo Mosso built the first device to measure brain blood flow: a delicately balanced bed a person lay on. Its head end dipped when the person began to think or feel something, as though mental effort had drawn blood to the brain.
Mosso’s manuscripts were rediscovered and analysed only recently. They show he had already grasped modern functional imaging’s core idea: that local blood flow tracks local neural activity. He even anticipated problems like background noise and the need for a well-designed task (Sandrone et al., 2014).
This blood-flow principle is the basis for modern fMRI scans.
Pneumoencephalography (1918)
This brutal-sounding method involved draining cerebrospinal fluid and replacing it with air to make brain structures visible on X-rays.
While it helped visualize the brain’s ventricles, it caused severe side effects like headaches, nausea, and discomfort. Still, it marked a major step toward visualizing internal brain anatomy.
Cerebral Angiography (1927)
Developed by Egas Moniz, this method used contrast dyes injected into the bloodstream to make brain blood vessels visible on X-rays.
Though invasive, it was revolutionary for diagnosing brain tumors and vascular conditions. This technique is still used today in a safer, refined form.
The Rise of Modern Imaging (1970s)
The invention of Computed Tomography (CT) in the 1970s brought non-invasive imaging into clinical use.
Shortly after, Magnetic Resonance Imaging (MRI) was developed, offering clear, detailed images of brain tissue without radiation. These advances ushered in a new era of safe, high-resolution brain scanning.
What Can Neuroimaging Reveal?
Neuroimaging can help identify:
- Brain structure issues: tumors, strokes, trauma.
- Mental health patterns: activity differences in depression, anxiety, or schizophrenia.
- Brain function: which areas light up during reading, math, or emotion regulation.
It’s a powerful tool for both clinical diagnosis and understanding the biology behind behavior.
Can Brain Scans Read Thoughts?
Not quite. While functional scans show which areas of the brain are active, they can’t reveal exact thoughts or feelings.
For example, seeing activity in the amygdala may suggest emotional arousal, but it doesn’t tell us what the person is feeling or why.
How Neuroimaging Helps in Psychology
Researchers use brain scans to:
- Test theories of behavior and emotion.
- Compare brain differences in mental health conditions.
- Track changes after therapy or medication.
For instance, studies show people with ADHD often have different patterns of activation in attention-related brain regions. Neuroimaging can help tailor treatments based on individual brain responses.
Critical Evaluation of Neuroimaging
Brain imaging is a genuinely contested method, and reading its evidence well means knowing where it can mislead.
- Reverse inference: assuming a brain region’s activity proves a specific thought, just because that region is “known” for that process.
- Correlational, not causal: seeing a region light up shows activity, not that the region causes or is needed for a mental process.
- The multiple-comparisons problem: testing thousands of brain locations at once means some will look “active” by chance alone unless properly corrected.
- Costly: techniques like fMRI and PET are expensive to run and analyse.
- Interpretation required: scan data must be analysed carefully by experts, not read directly off the image.
- Ethical concerns: misuse could lead to overgeneralising, such as using a scan to predict behaviour.
Reverse Inference
Reverse inference is a common but flawed way of reading a scan. A researcher sees that a brain region is active during a task and recalls that the region is “known” to support a certain mental process. The researcher then concludes the person must have been doing that process.
This is the reverse-inference fallacy.
The problem is that most brain regions take part in many different mental processes, not just one. Whether the inference is safe depends on how selectively a region responds to that one process. For most regions, it responds to too many others for the logic to hold (Poldrack, 2006).
Reverse inference can still be a useful starting hunch. Treated as proof, though, it produces confident nonsense: a region “lighting up” does not, by itself, tell you what someone was thinking. It is a hunch, not evidence.
Correlational, Not Causal
A brain scan is fundamentally correlational. It shows that activity in a region co-occurs with a mental state, not that the region causes it or is required for it.
A region might light up incidentally, as a downstream effect, or simply because the person is generally engaged with the task. Showing that a region is truly necessary for a function needs different evidence: lesion studies, or techniques that temporarily disrupt a region and observe what changes.
One shows correlation, the other can show cause. This is why cognitive neuroscience treats scanning and lesion research as complementary, not interchangeable.
Imaging findings and interference findings do different jobs. Scans reveal where activity correlates with a mental state. Lesion and stimulation studies reveal what a region actually causes.
The Multiple-Comparisons Problem
A whole-brain fMRI analysis can test tens of thousands of brain locations at once. Purely by chance, some of them will cross a statistical threshold and look “significant” unless the analysis corrects for how many tests were run.
One study makes this unforgettable.
- Aim: To show how uncorrected statistics can manufacture “brain activity” from pure noise.
- Method: Researchers placed a dead Atlantic salmon in an fMRI scanner and showed it photographs meant to trigger an emotional response. They analysed the scan as they would for a live participant, without correcting for multiple comparisons.
- Results: The analysis found statistically significant “activation” in the dead salmon’s brain.
- Conclusion: Uncorrected whole-brain statistics can manufacture convincing-looking activity from noise alone, so correcting for multiple comparisons is essential (Bennett et al., 2009).
The salmon had no living brain activity to detect, which is exactly the point. The dead salmon result triggered a lasting shift toward more conservative statistical thresholds in fMRI research. It remains the field’s standard cautionary tale about testing thousands of brain locations without correction.
Contemporary Research
Since 2015, neuroimaging has turned its sharpest tools on itself, asking whether its own findings can be trusted. Three studies, read together, tell a connected story.
First, a 2016 re-analysis found that common fMRI software produced false positives far above the claimed rate. In some cases, the false-positive rate reached 70% (Eklund et al., 2016). Then came the most direct test yet of whether fMRI results depend on who analyses them.
- Aim: To measure how much an fMRI study’s conclusions depend on the analyst’s own choices rather than on the data itself.
- Method: The same raw fMRI dataset, from participants doing a gambling task, was given to 70 independent research teams. Each team tested the same nine hypotheses using whatever analysis pipeline they judged best.
- Results: No two teams used the same pipeline, and their conclusions often split close to evenly between “significant” and “not significant” for the same hypothesis.
- Conclusion: A single fMRI dataset can support meaningfully different conclusions depending on legitimate, ordinary analysis choices, not researcher error (Botvinik-Nezer et al., 2020).
The reform movement did not stop there. Correlating brain measurements with traits such as intelligence turns out to need far larger samples than most studies have used. The typical study of a few dozen people is too small for its results to replicate reliably (Marek et al., 2022).
Together, these findings point to one fix: pre-register analyses, share data and code, and run larger studies. Neuroimaging is not becoming less useful. It is becoming harder to fool.
What’s Next for Neuroimaging?
Advancements on the horizon include:
- Real-time brain-computer interfaces.
- Portable neuroimaging devices.
- AI-enhanced image analysis.
- Personalized treatments using brain scan profiles.
As technology evolves, neuroimaging could help us better predict, prevent, and treat brain-based conditions.
FAQs
Can neuroimaging diagnose mental illness?
It can support a diagnosis, but it’s not a standalone tool. Diagnosis still relies on clinical interviews and behavioral assessments.
Is neuroimaging safe for children?
Yes—non-invasive techniques like MRI and EEG are safe and commonly used with children.
What’s the difference between MRI and CT?
MRI uses magnets and shows more detail. CT uses X-rays and is faster but exposes you to radiation.
Does neuroimaging show emotions?
It shows the brain regions involved in emotion but not the specific feeling itself.
Key Takeaways
- What It Is: Neuroimaging is a family of techniques that produce images of the brain’s structure or activity without opening the skull.
- Structure vs Function: Structural scans (MRI, CT) show brain anatomy; functional scans (fMRI, PET, EEG, MEG) show brain activity.
- Resolution Trade-off: fMRI and PET localise activity precisely but track it slowly; EEG and MEG track timing precisely but localise it only roughly.
- Correlational, Not Causal: A scan shows that brain activity co-occurs with a mental state, not that the activity causes it.
- Reform Movement: Since 2015, large studies have shown that some fMRI methods produced too many false positives. This pushed the field toward larger samples and more transparent methods.
- Not Mind-Reading: A brain scan cannot read a specific thought; it can only show that a region is more or less active.
References
American Psychological Association. (2014, August). Scanning the Brain. https://www.apa.org/action/resources/research-in-action/scan
Artico, M., Spoletini, M., Fumagalli, L., Biagioni, F., Ryskalin, L., Fornai, F., Salvati, M., Frati, A., Pastore, F. S., & Taurone, S. (2017). Egas Moniz: 90 Years (1927-2017) from Cerebral Angiography. Frontiers in neuroanatomy, 11, 81.
Bennett, C. M., Baird, A. A., Miller, M. B., & Wolford, G. L. (2009). Neural correlates of interspecies perspective taking in the post-mortem Atlantic salmon: An argument for proper multiple comparisons correction. Journal of Serendipitous and Unexpected Results, 1(1), 1–5.
Botvinik-Nezer, R., Holzmeister, F., Camerer, C. F., Dreber, A., Huber, J., Johannesson, M., … Schonberg, T. (2020). Variability in the analysis of a single neuroimaging dataset by many teams. Nature, 582(7810), 84–88. https://doi.org/10.1038/s41586-020-2314-9
Britannica, T. Editors of Encyclopaedia (2009, February 9). Pneumoencephalography. Encyclopedia Britannica. https://www.britannica.com/science/pneumoencephalography
Carter, R. (2019). The Human Brain Book: An Illustrated Guide to its Structure, Function, and Disorders (3rd ed). DK.
Eklund, A., Nichols, T. E., & Knutsson, H. (2016). Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates. Proceedings of the National Academy of Sciences, 113(28), 7900–7905. https://doi.org/10.1073/pnas.1602413113
Grant, F. C. (1927). Indications for and technic of ventriculography. Radiology, 9 (5), 388-395.
Lumen (n.d.). Brain Imaging Techniques. Retrieved June 8, 2021, from https://courses.lumenlearning.com/boundless-psychology/chapter/brain-imaging-techniques/
Marek, S., Tervo-Clemmens, B., Calabro, F. J., Montez, D. F., Kay, B. P., Hatoum, A. S., … Dosenbach, N. U. F. (2022). Reproducible brain-wide association studies require thousands of individuals. Nature, 603(7902), 654–660. https://doi.org/10.1038/s41586-022-04692-3
Ogawa, S., Lee, T. M., Nayak, A. S., & Glynn, P. (1990). Oxygenation-sensitive contrast in magnetic resonance image of rodent brain at high magnetic fields. Magnetic Resonance in Medicine, 14(1), 68–78. https://doi.org/10.1002/mrm.1910140108
Poldrack, R. A. (2006). Can cognitive processes be inferred from neuroimaging data? Trends in Cognitive Sciences, 10(2), 59–63. https://doi.org/10.1016/j.tics.2005.12.004
Sandrone, S., Bacigaluppi, M., Galloni, M. R., Cappa, S. F., Moro, A., Catani, M., Filippi, M., Monti, M. M., Perani, D. & Martino, G. (2014). Weighing brain activity with the balance: Angelo Mosso’s original manuscripts come to light. Brain, 137 (2), 621-633.
Van Wyhe, J. (2000). The History of Phrenology. Victorian Web. https://victorianweb.org/science/phrenology/intro.html
Further Reading
- Kniermin J. Neuroscience online: an electronic textbook for the neurosciences. Chapter 5: Cerebellum. University of Texas Health Science Center at Houston.
- Electroencephalography (EEG)
- Casson, A. J., Abdulaal, M., Dulabh, M., Kohli, S., Krachunov, S., & Trimble, E. (2018). Electroencephalogram. In Seamless Healthcare Monitoring (pp. 45-81). Springer, Cham.