Functional MRI
Functional MRI (fMRI) maps brain activity by measuring small changes in blood oxygenation that follow neural activity.
The usual signal is blood oxygen level dependent contrast: active cortex receives more oxygenated blood than it uses, so the local concentration of deoxyhaemoglobin falls and the T2* signal rises by 1 to 5%.
The study uses a 3 tesla scanner (1.5 tesla is possible) with a head coil and fast echo planar imaging.
Images covering the whole brain are taken every 1 to 3 seconds while the patient performs tasks in blocks of 20 to 30 seconds, alternating with rest. A high resolution T1-weighted anatomical scan is added for overlay.
Resting state fMRI needs no task. The patient lies still for 6 to 10 minutes, and spontaneous fluctuations in the signal reveal networks of regions that act together (motor, language, visual, default mode and others). This suits children, sedated and uncooperative patients.
The examination takes 30 to 60 minutes. There is no ionising radiation, and no contrast is needed.
fMRI is used clinically mainly before neurosurgery. Brain tumours, epilepsy foci and vascular malformations near eloquent cortex (motor, sensory, language, visual) carry a risk of a permanent deficit. Mapping lets the surgeon plan the approach, the extent of resection and the need for awake surgery.
In adults, the motor cortex is localised with a sensitivity of about 90% and the dominant language hemisphere with a concordance of about 90% with invasive tests. Language localisation within the hemisphere is less reliable than motor mapping.
Indications.
- Pre-operative mapping of motor, sensory, language and visual cortex near a tumour, vascular malformation or epileptic focus.
- Assessment of language dominance (an alternative to the intracarotid amobarbital test) before temporal lobe or other epilepsy surgery.
- Planning of radiotherapy or radiosurgery to spare eloquent areas.
- Assessment of cerebral reorganisation after stroke, tumour growth or early injury.
- Disorders of consciousness: detection of covert awareness in patients with vegetative or minimally conscious state (specialist research and tertiary use).
- Resting state networks in epilepsy, dementia, psychiatric disease and developmental disorders (mostly research).
- Pain, memory and cognitive studies in research settings.
Not for routine diagnosis of neurological or psychiatric disease. fMRI is not a test for lie detection or for individual level diagnosis of mental illness.
Contraindications and precautions.
- The usual MRI safety screen: pacemakers (unless conditional), non-conditional aneurysm clips, cochlear implants and metallic foreign bodies.
- Claustrophobia, inability to cooperate with tasks, and excess head motion (more than 2 to 3 mm) make data unusable.
- Gadolinium is not needed for the functional scan, and is used only for the anatomical study if the tumour needs it.
- Anti-epileptic drugs, sedatives and caffeine alter the signal. Sedation prevents most task studies.
- Pregnancy: acceptable if essential, as with all MRI.
- Patients with severe weakness, aphasia or visual loss may be unable to do the tasks.
Preparation and technique.
The patient is trained in the tasks before scanning, outside the scanner. Head motion is reduced with padding and a head immobiliser, and earplugs protect the hearing from the loud echo planar noise. Visual stimuli are shown on a screen or goggles, and sound is presented through headphones.
Typical tasks are finger tapping, hand squeezing, foot or toe movement and tongue movement for motor mapping, touch for sensory cortex, verb generation, word fluency, naming and sentence listening for language, and flashing checkerboards for vision. A task is chosen for the patient's ability and the lesion site.
Each task is repeated in several blocks, and a rest period follows. The data are corrected for motion, smoothed and analysed statistically to produce maps of significance overlaid on the anatomical scan. Thresholds are chosen with the surgeon, because they change the apparent size of the active area.
