MRI Lesion Does Not Always Mean Seizure Cause

Finding a lesion on MRI can feel like finding the answer.

Sometimes it is. A focal cortical dysplasia, hippocampal sclerosis, low-grade epilepsy-associated tumour, cavernoma, stroke scar or other structural abnormality may fit the seizure type, EEG and clinical history so well that it becomes a compelling explanation for the epilepsy.

But an MRI abnormality is not automatically the place where seizures begin.

Some findings are incidental. Some are related to epilepsy but not independently responsible for every seizure. Some people have more than one abnormality. And sometimes the lesion that looks most obvious is not the most important epileptogenic structure.

MRI shows structure, not seizure onset

Structural MRI is designed to show anatomy.

It can reveal scars, malformations, tumours, vascular abnormalities, atrophy and many other changes in the brain.

What it does not directly show is an epileptic seizure beginning.

That distinction matters. A structural lesion can be highly epileptogenic, mildly epileptogenic, unrelated to the person’s habitual seizures, or simply an incidental finding discovered because the brain was scanned.

Incidental brain findings are real

Brain MRI often finds abnormalities in people who have no neurological symptoms at all.

A systematic review of 19,559 neurologically asymptomatic people found incidental neoplastic findings in about 0.7% and non-neoplastic findings in about 2%, excluding common white-matter hyperintensities, silent infarcts and microbleeds.

More recent systematic reviews have confirmed that incidental intracranial findings remain common as MRI resolution improves.

This means a lesion found during an epilepsy investigation cannot automatically be assumed to explain the seizures merely because it exists.

The more detailed the scan, the more we may find

High-resolution MRI improves detection of subtle epileptogenic lesions — which is exactly why dedicated epilepsy protocols are so useful.

But greater sensitivity also increases the likelihood of finding abnormalities whose clinical significance is uncertain.

The question therefore changes from ‘Is there anything abnormal?’ to ‘Does this abnormality fit the epilepsy?’

See Why a Dedicated Epilepsy MRI Is Different for how epilepsy-specific imaging improves lesion detection.

What does concordance mean?

Concordance means that different pieces of evidence point toward the same brain region or network.

For example, a person may have:

• seizure symptoms suggesting the left temporal lobe;
• interictal or ictal EEG abnormalities maximal over the left temporal region;
• left hippocampal sclerosis on MRI;
• and neuropsychological findings compatible with dysfunction in that same network.

When independent tests agree, confidence that the MRI lesion is clinically relevant becomes much stronger.

Discordance is a warning to investigate further

Discordance means the pieces do not line up.

An MRI lesion may be on the right while the habitual seizure semiology and EEG repeatedly point left.

The EEG may appear multifocal despite one visible lesion.

The seizure symptoms may suggest a region anatomically distant from the MRI abnormality.

None of these automatically proves the lesion is incidental, but they make a simple lesion-equals-seizure-source assumption unsafe.

A lesion can be real and still not be the whole story

Some abnormalities are genuinely associated with epilepsy but sit within a wider epileptic network.

For example, seizure activity may arise from the lesion itself, cortex immediately around it, or a connected network that has become chronically epileptogenic.

In other cases, the lesion may have helped create the network but may not correspond exactly to the earliest measurable seizure onset.

This is one reason epilepsy surgery planning is more sophisticated than drawing a line around an MRI abnormality.

The epileptogenic zone is not identical to the MRI lesion

Epilepsy surgery uses the concept of the epileptogenic zone: the area of cortex that must be removed, disconnected or otherwise treated to achieve seizure freedom.

The MRI-visible lesion is one piece of evidence used to estimate that zone.

The lesion may be smaller than the epileptogenic zone, larger than it, overlap it closely, or occasionally lie outside the true seizure-onset network.

The distinction becomes especially important when surgery would involve language, memory, motor or sensory cortex.

Lesion does not always equal seizure-onset zone

A tumour is a good example.

Some epilepsy-associated tumours are strongly linked with seizures, but seizure generation can involve surrounding cortex as well as the tumour itself.

The same principle applies to vascular malformations and some malformations of cortical development.

Removing only what looks abnormal on MRI is therefore not always equivalent to treating the full epileptogenic network.

Hippocampal sclerosis is strongly associated with temporal lobe epilepsy — but context still matters

Hippocampal sclerosis is one of the classic MRI abnormalities associated with temporal lobe epilepsy.

When seizure semiology and EEG agree with the side of the hippocampal abnormality, the finding may be highly meaningful.

But hippocampal structural variants can also occur, and hippocampal sclerosis may coexist with another lesion.

The site has a separate page on Hippocampal Sclerosis.

Dual pathology

Dual pathology describes situations in which hippocampal sclerosis coexists with a second principal brain lesion such as a tumour, vascular malformation, scar or malformation of cortical development.

Neuropathology literature also distinguishes this from ‘double pathology’, where two principal lesions occur without hippocampal sclerosis.

The important clinical point is that one person can have more than one potentially epileptogenic abnormality.

Finding one lesion does not guarantee that the search is finished.

Why dual pathology matters for surgery

If two abnormalities contribute to the epileptogenic network, treating only one may leave the other capable of generating seizures.

Conversely, one lesion may be incidental while the other is the true driver.

Presurgical teams therefore examine whether each abnormality matches the electroclinical data rather than assuming that the most visually obvious lesion is the most important one.

What about white-matter changes?

Small white-matter hyperintensities become increasingly common with age and vascular risk.

They may reflect small-vessel disease, migraine-related changes or other processes.

Many are not epileptogenic lesions.

A report saying ‘nonspecific white-matter change’ should therefore not automatically be translated into ‘this is what causes the seizures’.

Arachnoid cysts are another classic example

Arachnoid cysts are frequently discovered incidentally on brain imaging.

Some are large or strategically located and may be clinically relevant, but many remain asymptomatic.

If a person with seizures has an arachnoid cyst, the team still has to ask whether the electroclinical evidence points to cortex adjacent to that cyst or somewhere else entirely.

What about Chiari malformation?

Chiari malformations are also found incidentally in some people undergoing MRI.

They can cause specific neurological symptoms, but they are not automatically explanations for epilepsy.

The presence of a Chiari malformation in somebody who also has seizures therefore requires the same discipline: association is not proof of causation.

Meningiomas and other tumours

Some brain tumours can unquestionably cause seizures, particularly when they involve or irritate cerebral cortex.

But small incidental meningiomas are also detected in people who have never had a seizure.

Location, surrounding oedema, cortical involvement, seizure semiology and EEG all influence whether the tumour is considered likely to be epileptogenic.

Cavernomas

Cerebral cavernous malformations can cause epilepsy, haemorrhage or focal neurological symptoms.

But not every cavernoma found on MRI is necessarily responsible for a person’s seizures.

Multiple cavernomas can make localisation even more complicated because several lesions may be present while only one is clinically active.

This is one reason the planned Structural & Acquired section needs a dedicated cavernoma page rather than treating every vascular lesion as equivalent.

Focal cortical dysplasia can be subtle — or more extensive than it looks

Focal cortical dysplasia (FCD) is a major cause of drug-resistant focal epilepsy.

Some FCDs are obvious on MRI. Others are extremely subtle or initially MRI-negative.

Even when visible, the anatomical abnormality may not define the full epileptogenic network.

For more detail, see Focal Cortical Dysplasia.

A visible lesion can bias interpretation

Once clinicians know an MRI lesion exists, there is a human tendency to interpret later information through that finding.

This is a form of anchoring bias.

Epilepsy centres try to reduce that risk by integrating semiology, electrophysiology, imaging and neuropsychology rather than allowing any single test to dominate automatically.

Multidisciplinary case review is therefore not bureaucracy; it is a safeguard against premature conclusions.

The seizure description still matters enormously

Seizure semiology — what happens before, during and immediately after an event — provides anatomical information.

A stereotyped visual aura, unilateral tonic posture, language disturbance, autonomic symptom or experiential phenomenon can suggest which networks are being recruited.

If that pattern repeatedly disagrees with the MRI lesion, clinicians may need to investigate whether the lesion is incidental or whether seizure propagation is creating a misleading outward picture.

Scalp EEG adds another independent line of evidence

EEG can show interictal epileptiform discharges or record ictal onset patterns.

When these correspond with an MRI lesion, confidence rises.

When they repeatedly point elsewhere, the case becomes more complex.

EEG itself has limitations, especially for deep or rapidly propagating seizures, so discordance is a reason for further assessment rather than an automatic veto.

PET, SPECT and MEG can help when MRI and EEG disagree

Functional imaging and source-localisation tools may be used in specialist presurgical evaluation when standard tests do not provide a coherent answer.

FDG-PET can identify regions of altered metabolism.

Ictal SPECT can show perfusion changes associated with a seizure.

MEG can help localise interictal magnetic fields.

These tests do not replace the MRI lesion; they add independent information about function and network localisation.

See PET, SPECT and MEG in Epilepsy.

Sometimes intracranial EEG is needed

If non-invasive data remain discordant or if the suspected seizure-onset zone lies near eloquent cortex, intracranial EEG may be considered.

SEEG can sample structures in three dimensions and test competing hypotheses about where seizures begin and how they spread.

The crucial word is hypothesis: electrodes are placed to answer a specific localisation question, not simply because an MRI lesion exists.

See SEEG: Recording Seizures From Inside the Brain.

What if MRI shows nothing?

The mirror-image error is assuming that normal MRI means there is no structural basis for epilepsy.

Some epileptogenic abnormalities are too subtle for routine imaging, are missed because the protocol was not epilepsy-specific, or become apparent only after expert re-review or post-processing.

That topic is covered separately in Myth: A Normal MRI Means There Is Nothing Causing the Epilepsy.

What if the lesion and seizure side disagree?

This is one of the situations that most strongly pushes an epilepsy team toward further investigation.

Possible explanations include rapid seizure propagation, incorrect scalp EEG lateralisation, a lesion that is not epileptogenic, a second MRI-occult lesion, bilateral abnormalities, or a wider epileptic network.

The correct answer cannot usually be chosen from MRI alone.

Concordance is particularly important before surgery

The ILAE Neuroimaging Task Force emphasises that structural MRI should be optimally acquired and interpreted in the context of the full clinical data.

When a clear MRI lesion is congruent with the electrophysiological abnormality, surgical planning may be more straightforward and outcomes can be better.

When data are discordant, specialist investigation becomes more important because treating the wrong region can fail to control seizures while exposing the person to unnecessary neurological risk.

Why ‘lesional epilepsy’ is not a complete answer

The term lesional epilepsy can be useful shorthand, but it can hide several questions:

• Is the lesion actually epileptogenic?
• Does seizure onset occur within the lesion or nearby cortex?
• Is there a second lesion?
• Is the epileptogenic network larger than the visible lesion?
• Does the lesion overlap critical functional cortex?
• Would treating the lesion alone be sufficient?

Those are localisation questions, not merely radiology questions.

The practical rule

An MRI abnormality can be hugely important.

But the correct diagnostic reasoning is not:

‘There is a lesion, therefore that lesion causes the epilepsy.’

It is:

‘There is a lesion. Does the rest of the evidence show that this lesion belongs to the epileptogenic network responsible for the person’s habitual seizures?’

That is why epilepsy diagnosis and surgical planning combine structure, electrical activity, clinical semiology and — when necessary — functional imaging and intracranial recordings.

Related pages on The Untold Side

MRI and Neuroimaging in Epilepsy
Why a Dedicated Epilepsy MRI Is Different
A Normal MRI Does Not Mean There Is No Cause
Focal Cortical Dysplasia
Hippocampal Sclerosis
Epilepsy Surgery and Presurgical Assessment
SEEG

Sources and further reading

Bernasconi A et al. Recommendations for the use of structural MRI in the care of patients with epilepsy. ILAE Neuroimaging Task Force. Epilepsia. 2019.
Bernasconi A et al. MRI essentials in epileptology: a review from the ILAE Imaging Taskforce. Epileptic Disorders. 2020.
Morris Z et al. Incidental findings on brain MRI: systematic review and meta-analysis. BMJ. 2009.
Prevalence of incidental intracranial findings on MRI: systematic review and meta-analysis. 2022.
Surgical pathology of epilepsy-associated non-neoplastic cerebral lesions: hippocampal sclerosis, focal cortical dysplasia and dual pathology.
ILAE Commission on Diagnostic Methods. International recommendation for comprehensive neuropathologic workup of epilepsy surgery tissue. Epilepsia. 2016.

Evidence checked: September 2026.