Functional MRI (fMRI) maps changes in blood oxygenation associated with brain activity. In epilepsy surgery, it is used mainly to identify language, motor and other eloquent networks that surgeons need to preserve.
What fMRI measures
Most clinical fMRI uses the BOLD signal—blood-oxygen-level-dependent contrast.
It does not record neurons directly. It measures vascular changes coupled to neuronal activity.
Why language mapping matters
Language is not organised identically in every person.
Long-standing epilepsy can also contribute to developmental or acquired reorganisation of language networks, making individual mapping especially important.
Language lateralisation
One major question is which hemisphere is dominant for language.
Task-based fMRI can help determine whether language is predominantly left-sided, right-sided or bilateral.
Language localisation
Knowing the dominant hemisphere is not the same as knowing exactly which cortical tissue is essential for naming, comprehension or speech production.
fMRI can show activated language networks, but direct electrical stimulation may still be needed when surgery approaches critical language cortex.
Motor mapping
Motor tasks can map hand, foot or face-related motor regions.
Motor mapping is generally more anatomically reproducible than language mapping.
Visual and sensory mapping
Some specialist centres use fMRI to map visual or sensory networks when surgical plans approach eloquent cortex.
Task performance matters
Task-based fMRI depends on the person understanding and performing the task consistently.
Young children, people with intellectual disability, severe language impairment or anxiety may produce less reliable results.
Resting-state fMRI
Resting-state methods analyse spontaneous network activity without requiring a task.
They are increasingly studied for people who cannot reliably perform conventional tasks, but protocols and clinical interpretation remain less standardised.
fMRI and the Wada test
fMRI has replaced the Wada test for language lateralisation in many epilepsy centres.
However, the 2026 clinical review notes that Wada testing can remain useful in selected cases, particularly where there is concern about bilateral memory risk or discordant non-invasive findings.
Memory fMRI
Memory paradigms may help estimate risk of postoperative memory decline, especially before temporal lobe surgery.
They are interpreted alongside neuropsychological testing, structural MRI and other presurgical data.
Comparison with electrical stimulation mapping
A meta-analysis comparing language fMRI with electrical stimulation showed moderate pooled sensitivity and specificity but substantial heterogeneity.
fMRI therefore cannot simply replace direct stimulation in every case where exact language cortex must be defined.
False activation and false absence
Movement, poor task performance, vascular abnormalities and statistical thresholds can all affect fMRI maps.
An area that does not activate is not automatically safe to remove.
Worldwide access
Clinical fMRI requires appropriate MRI hardware, validated paradigms, analysis software and specialised interpretation.
Language-specific tasks also need cultural and linguistic adaptation, making worldwide standardisation difficult.
The central point
fMRI maps functional networks to reduce surgical risk; it does not identify the seizure focus by itself. Its strongest established role is presurgical language lateralisation and functional mapping.
Related reading
Epilepsy Surgery and Presurgical Assessment →
Neuropsychology in Presurgical Epilepsy Assessment →
Sources
- Clinical Functional MRI in Epilepsy — 2026 review
- 2024 recommendations on language-lateralisation fMRI in epilepsy
- Meta-analysis comparing fMRI and electrical stimulation for language localisation
Evidence checked: September 2026.