Functional brain imaging plays an important role in seizure focus localisation. However, truly ictal single-photon emission tomography (SPET) studies are not routinely performed owing to technical problems associated with the use of tracers and methodological and logistical difficulties. In this study we tried to resolve both of these issues by means of a new procedure: technetium-99m ethyl cysteinate dimer (ECD) brain SPET performed during seizures pharmacologically provoked with pentylenetetrazol, a well-known central and respiratory stimulant. We studied 33 drug-resistant epileptic patients. All patients underwent anamnestic evaluation, neuropsychological and psychodynamic assessment, magnetic resonance imaging, interictal and ictal video-EEG monitoring, and interictal and ictal SPET with (99m)Tc-ECD. In order to obtain truly ictal SPET, 65 mg of pentylenetetrazol was injected every 2 minutes and, immediately the seizure began, 740 MBq of (99m)Tc-ECD was injected. The scintigraphic findings were considered abnormal if a single area of hyperperfusion was present and corresponded to the site of a single area of hypoperfusion at interictal SPET: the "hypo-hyperperfusion" SPET pattern. In 27 of the 33 patients (82%), interictal-ictal SPET showed the hypo-hyperperfusion SPET pattern. Video-EEG showed a single epileptogenic zone in 21/33 patients (64%), and MRI showed anatomical lesions in 19/33 patients (57%). Twenty-two of the 27 patients with hypo-hyperperfusion SPET pattern underwent ablative or palliative surgery and were seizure-free at 3 years of follow-up. No adverse effects were noted during pharmacologically provoked seizure. It is concluded that ictal brain SPET performed during pharmacologically provoked seizure provides truly ictal images because (99m)Tc-ECD is injected immediately upon seizure onset. Using this feasible procedure it is possible to localise the focus, to avoid the limitations due to the unpredictability of seizures, to avoid pitfalls due to late injection, to avoid intracranial EEG recording and to minimise costs. The clinical value of our method is confirmed by the good outcome after 3 years of follow-up in those patients submitted to ablative or palliative surgery.