Step 1: Frame the clinical problem.
This boy has drug-resistant epilepsy, meaning seizures continue despite 4 well-chosen antiepileptic drugs, so he is being worked up for epilepsy surgery. His childhood febrile illness raises suspicion of mesial temporal sclerosis, but the epilepsy-protocol MRI is normal, so the usual structural clue is missing.
Step 2: State the goal of presurgical evaluation.
Before surgery, the seizure onset zone must be localized with confidence. When MRI is normal, this is called MRI-negative or non-lesional epilepsy, and localization then depends much more on functional and electro-clinical data.
Step 3: Compare the non-invasive tools.
Interictal scalp EEG, done between seizures, only shows interictal spikes and cannot reliably pinpoint the seizure onset zone by itself. Video EEG alone records the clinical seizure and its ictal EEG pattern, which is essential for confirming the diagnosis and seizure semiology, but scalp EEG signal can be too diffuse to localize a deep or non-lesional focus precisely. Interictal FDG-PET shows a zone of reduced glucose metabolism between seizures, which is helpful but often shows a wider area of hypometabolism than the true onset zone.
Step 4: Explain why ictal SPECT combined with video EEG is best.
During a seizure, blood flow rises sharply at the true seizure onset zone within seconds. Injecting the HMPAO tracer as close to seizure onset as possible and imaging afterward captures this ictal hyperperfusion, which is far more focal and reliable than interictal PET hypometabolism. Pairing this with simultaneous video EEG gives both the precise electro-clinical onset time and the functional localization, which is the standard non-invasive strategy in MRI-negative drug-resistant epilepsy.
Step 5: Final answer.
\[ \boxed{\text{Video EEG with ictal 99mTc-HMPAO brain SPECT}} \]