Understanding drug-resistant epilepsy reports: EEG findings, MRI lesions, and the final presurgical opinion

It can seem contradictory when an EEG describes discharges, an MRI is reported as normal, and the clinician still recommends further assessment. The tests may be answering different questions. EEG describes electrical activity, imaging examines structure or function, and the diagnosis also depends on the actual events. Start by identifying what each report establishes before deciding how much it changes the treatment discussion. NICE NG217: Diagnosis and assessment of epilepsyBernasconi et al.: ILAE structural MRI recommendations and HARNESS protocol, 2019

Key takeaways

These excerpts come from the original article. Read the full sections below for context.

An EEG report generally includes recording conditions, observed findings, an impression, and clinical interpretation. Age, wakefulness or sleep, medicines, recent seizures, and recording quality matter. A cropped final sentence saying abnormal EEG removes much of the context needed to understand the result. Kane et al.: IFCN revised EEG glossary and report format, 2017
Hippocampal sclerosis and focal cortical dysplasia can be structural explanations associated with epilepsy. Their relevance to the current events depends on agreement with clinical and electrical evidence. The name of a structural abnormality does not by itself mean malignant cancer or establish the indication for a particular operation. ILAE EpilepsyDiagnosis: Structural etiologies and imaging
Submit complete reports and source data through a center offering the relevant services. State medicines, events, and clinical circumstances at the time of each study. CAAE information can help identify an inquiry route, but confirm access to EEG review, imaging expertise, and neuropsychological assessment with the actual hospital. 中国抗癫痫协会：第五批一、二级癫痫中心评审结果，2025年9月28日

Quick answer

It can seem contradictory when an EEG describes discharges, an MRI is reported as normal, and the clinician still recommends further assessment. The tests may be answering different questions. EEG describes electrical activity, imaging examines structure or function, and the diagnosis also depends on the actual events. Start by identifying what each report establishes before deciding how much it changes the treatment discussion. NICE NG217: Diagnosis and assessment of epilepsy Bernasconi et al.: ILAE structural MRI recommendations and HARNESS protocol, 2019

Full guide

It can seem contradictory when an EEG describes discharges, an MRI is reported as normal, and the clinician still recommends further assessment. The tests may be answering different questions. EEG describes electrical activity, imaging examines structure or function, and the diagnosis also depends on the actual events. Start by identifying what each report establishes before deciding how much it changes the treatment discussion. NICE NG217: Diagnosis and assessment of epilepsy Bernasconi et al.: ILAE structural MRI recommendations and HARNESS protocol, 2019

Separate the description from its interpretation

An EEG report generally includes recording conditions, observed findings, an impression, and clinical interpretation. Age, wakefulness or sleep, medicines, recent seizures, and recording quality matter. A cropped final sentence saying abnormal EEG removes much of the context needed to understand the result. Kane et al.: IFCN revised EEG glossary and report format, 2017

Mark the observed finding, the proposed explanation, and anything still requiring clinical correlation. Suggested, possible, and cannot exclude express different levels of certainty. When seeking review in China, retain the full report and available source recording. A translation should preserve these qualifiers rather than turn every statement into a confirmed diagnosis.

Spikes and sharp waves are not a direct count of clinical seizures

Spikes, sharp waves, and related complexes describe particular patterns. Their epileptiform significance depends on morphology, distribution, state, and background. Interictal discharges can support understanding of epilepsy, but their number cannot be converted directly into the number of daily seizures or a rate of ongoing brain injury. Kane et al.: IFCN revised EEG glossary and report format, 2017

If clinical control has improved but an EEG remains abnormal, ask whether the finding changes the risk assessment or plan. Do not increase treatment independently because of one tracing. Likewise, fewer discharges do not by themselves establish that medication can safely be withdrawn. The clinical course remains essential to that decision. NICE NG217: Full guideline, updated January 30, 2025

The relevant question is often how the electrical result fits the patient's actual experience. A report may provide useful support without resolving every episode. The clinician should explain that relationship rather than expecting the patient to infer it from technical terminology.

Slowing, background change, and artifact mean different things

Slowing or an altered background describes electrical activity whose meaning depends on age, region, alertness, medicines, and disease context. These terms are not interchangeable with epileptiform discharges. Eye movements, muscle activity, and technical factors can also introduce artifact that must be distinguished from cerebral signals. Kane et al.: IFCN revised EEG glossary and report format, 2017

A statement that part of the record is affected by artifact identifies a limitation; it does not automatically make the entire study useless. Ask which portions remained interpretable and whether further recording could answer the unresolved question. Comparing line thickness on a phone screen is not a reliable way to grade disease severity.

No recorded seizure does not always settle the diagnosis

If the usual event was not captured, the study did not obtain complete electroclinical information about that event. That can limit certainty without proving the event never occurs. Interpretation considers baseline frequency, monitoring duration, sleep, and other recording conditions. Tatum et al.: ILAE/IFCN minimum standards for inpatient long-term video-EEG, Epilepsia 2022

Even a captured behavioral change needs to be checked against the person's typical episodes. Some deep or spatially limited seizures have unclear scalp manifestations. ILAE material on focal cortical dysplasia notes that ictal scalp changes can be subtle or poorly localized. Absence of an obvious scalp ictal pattern therefore cannot be independently rewritten by a family as a definite non-epileptic diagnosis. ILAE EpilepsyDiagnosis: Focal cortical dysplasia, EEG findings, 2024

Ask whether the report addresses every event type or only the one observed. The next step may differ when a frequent minor event was recorded but the patient's less frequent disabling event was not. Keeping those distinctions visible prevents overinterpretation of an otherwise useful study.

A functional-seizure diagnosis needs a clear explanation

The diagnosis rests on clinical features and appropriate evidence, not simply normal tests. The AAN's 2025 guideline emphasizes typical-event assessment and evaluation for coexisting epilepsy. Ask which features support the conclusion, whether all event patterns have been explained, and how the proposed care addresses them. AAN: Management of Functional Seizures, clinical guideline summary, December 2025

Functional seizures are not deliberate performance and do not make the person's difficulties unimportant. If both functional and epileptic events occur, the record should help the patient and caregivers distinguish them. Medication decisions depend on the complete diagnosis and should not follow from independently reacting to one newly encountered word.

Left, right, and generalized are not simple surgical boundaries

Discharge distribution can provide lateralizing or localizing evidence, but scalp recordings reflect propagated activity and what the electrodes can detect. More frequent abnormalities on one side do not automatically identify a small region that can be safely removed. A focal seizure can also develop into bilateral tonic-clonic activity, so outward appearance alone does not determine origin. Beniczky et al.: Updated seizure classification, ILAE position paper, Epilepsia 2025

Some terminology changed in the 2025 ILAE classification, including the way consciousness is described. Ask the clinician to connect older terms to the patient's actual manifestations. Unknown whether focal or generalized is also different from unclassified. Neither should be translated indiscriminately as a seizure arising throughout the entire brain. ILAE: Updated classification of epileptic seizures, 2025

MRI negative describes an imaging result, not the final treatment decision

MRI negative usually means that a relevant structural abnormality was not identified with the available acquisition and interpretation. Sequences, resolution, movement, and reading expertise can influence detection of subtle lesions. Review of source images is therefore more informative than relying only on the last line of the report. Bernasconi et al.: ILAE structural MRI recommendations and HARNESS protocol, 2019

Some patients can still progress through an assessment using clinical, EEG, and additional information. ILAE referral recommendations support timely consideration in appropriate drug-resistant patients rather than stopping solely because there is no clear lesion. Additional tests, however, cannot guarantee that a safely treatable region will ultimately be identified. Jehi et al.: Timing of referral for epilepsy surgery evaluation, ILAE consensus, 2022

If another MRI is recommended, ask what would be different: the acquisition, the clinical hypothesis guiding review, or a change in the condition. That explanation distinguishes a justified reassessment from an assumption that every new consultation must repeat all previous work.

Hippocampal sclerosis and cortical dysplasia are structural clues

Hippocampal sclerosis and focal cortical dysplasia can be structural explanations associated with epilepsy. Their relevance to the current events depends on agreement with clinical and electrical evidence. The name of a structural abnormality does not by itself mean malignant cancer or establish the indication for a particular operation. ILAE EpilepsyDiagnosis: Structural etiologies and imaging

Suggested dysplasia may still require specialist review and correlation. The visible boundary need not match the complete region relevant to seizure generation, and a patient should not draw their own resection map from the image. If institutions interpret the same study differently, ask whether the disagreement concerns quality, experience, or insufficient evidence.

PET hypometabolism is not an instruction to remove every abnormal region

In epilepsy assessment, PET commonly contributes metabolic information, while SPECT can provide perfusion information. Interpretation depends on seizure timing and technical conditions and must be related to the other findings. An abnormal region may represent a wider network rather than a uniform treatment target. Traub-Weidinger et al.: EANM PET and SPECT practice guideline for epilepsy, 2024

Ask which hypothesis the image supports, whether it conflicts with EEG or MRI, and how uncertainty would be tested. Reduced metabolism does not mean that the tissue is necessarily dead. It also cannot independently predict the amount of memory recovery or the exact duration of future seizure freedom.

Color scales and software displays can make a result look more definite than its clinical meaning. The written interpretation and the multidisciplinary discussion remain important. The value of a fusion image is its contribution to a coherent assessment, not its visual prominence.

Cognitive scores require personal context

Neuropsychological assessment provides a pretreatment reference and informs discussion of cognitive effects. Performance can be influenced by language, education, fatigue, mood, medicines, and the specific task. One low score should not automatically be equated with permanent intellectual decline or inevitable exclusion from surgery. Baxendale et al.: ILAE neuropsychological assessment in epilepsy surgery, 2019

For an international patient in China, the report should identify the testing language, reference population, and translation or communication limitations. Discuss whether results match everyday problems at home or work. Comparing before and after treatment needs appropriate methods and interpretation rather than merely subtracting two overall scores.

The patient's goals also matter. A small change in a particular function may have considerable meaning for someone's occupation or schooling. Ask the team to explain potential effects in terms of the activities the patient values, not only numerical categories.

Pathogenic, likely pathogenic, and uncertain genetic findings cannot be interchanged

Laboratories classify variants from available evidence, while the clinical team determines whether a finding explains the patient's events and other features. A variant of uncertain significance should not independently determine an irreversible treatment. A negative test means the method did not identify a reportable relevant result; it does not exclude every genetic mechanism. Krey et al.: Current practice in diagnostic genetic testing of the epilepsies, ILAE Genetics Commission, 2022

Keep the coverage, specimen, and laboratory interpretation. Family studies or later reanalysis may sometimes help. If an online account links a gene to a particular medicine, confirm the variant mechanism and clinical evidence. Different changes within the same gene can have different implications and should not be treated as one uniform drug indication.

Drug levels and laboratory flags are not stand-alone dose instructions

A medication concentration needs the sampling time, dose history, actual use, interactions, and clinical response. Some patients have seizures within a reference range, while others experience adverse effects at lower exposure. A population reference interval is not a personalized prescription for an effective and safe dose. NICE NG217: Full guideline, updated January 30, 2025

Liver, kidney, or blood-count abnormalities also need assessment in relation to the actual medicine and other possible causes. If significant symptoms develop, contact a clinician and describe them rather than sending an isolated laboratory image and waiting indefinitely. The 2026 medication guideline emphasizes differences between drugs and patient circumstances rather than a single adjustment formula. Wu et al.: Clinical practice guidelines for third-generation antiseizure medications, Seizure 2026;134:13–26

Keep the previous result and date when a trend is relevant. The rate and context of change may matter more than the presence of an arrow beside one value. Patients can organize that information without deciding the correct medical response themselves.

An antibody result must fit the clinical course

Some patients with recently worsening seizures and cognitive or behavioral change undergo assessment for autoimmune encephalitis. Blood or cerebrospinal-fluid antibodies provide only part of the evidence. Specimen, assay, phenotype, and alternative causes matter. An isolated or borderline positive result does not independently establish an indication for immune treatment. Abboud et al.: Autoimmune encephalitis diagnosis and acute management best-practice recommendations, 2021

Where the result and symptoms disagree, ask whether verification or another investigation is needed. Longstanding stable epilepsy and a rapidly developing encephalitic illness should not be assigned identical management merely because the same laboratory panel was ordered. Do not use an unsupervised steroid trial as a personal test of the diagnosis.

The multidisciplinary conclusion brings the evidence closest to a decision

A presurgical meeting may find the data concordant, identify a remaining localization question, or judge the risk greater than the likely benefit. Concordant usually means that several sources support a similar explanation; it does not imply zero operative risk. A recommendation for intracranial EEG is a further investigation, not a promise that resection will follow. Jehi et al.: Timing of referral for epilepsy surgery evaluation, ILAE consensus, 2022 NINDS: Epilepsy and Seizures

Ask about functions that must be protected, major risks, alternatives, and uncertainty. If resection is not appropriate now, other options may remain. Keep the reason alongside a statement that surgery is not currently recommended. Whether another assessment becomes worthwhile depends on new evidence rather than mechanically repeating an identical test list.

Make a second opinion in China a review of evidence, not just wording

Submit complete reports and source data through a center offering the relevant services. State medicines, events, and clinical circumstances at the time of each study. CAAE information can help identify an inquiry route, but confirm access to EEG review, imaging expertise, and neuropsychological assessment with the actual hospital. 中国抗癫痫协会：第五批一、二级癫痫中心评审结果，2025年9月28日

Prepare a one-page list of troubling terms and apparently conflicting conclusions. Ask the receiving clinician which findings change the assessment and which simply use different language. Request a written explanation that can be discussed with the home clinician. Report interpretation becomes useful when it clarifies the next decision rather than leaving the family to choose between unfamiliar technical words.