ANATOMIC CORRELATES OF MEMORY FROM INTRACAROTID AMOBARBITAL INJECTIONS WITH TECHNETIUM TC-99M HEXAMETHYLPROPYLENEAMINE OXIME SPECT

Objective.-To better identify regions of the brain affected by intracarotid amobarbital injections and to more precisely predict whether resections of specific brain regions will cause postoperative memory deficits. Design.-We modified the standard intracarotid amobarbital procedure by adding a radioactive tracer to the amobarbital injection, thereby providing better correlation between behavior and deactivated brain region. Setting.-Tertiary-care hospital center with a dedicated program for medical and surgical treatment of epilepsy. Patients.-We studied 39 patients with medically intractable epilepsy drawn from a regional referral base. Intervention.-Intracarotid injection of 125 mg of sodium amobarbital with 37 MBq of technetium Tc 99m hexamethylpropyleneamine oxime (HMPAO), followed by language and memory testing. Main Outcome Measures.-The distribution of amobarbital as measured by single photon emission computed tomographic imaging of HMPAO and patient performance on memory tasks. Results.-Medial temporal regions were irrigated by the amobarbital in only 28% of the injections. Overall, findings suggest that medial temporal and lateral neotemporal cortex play a role in memory. Conclusions.-The regions involved in memory function vary by individual, as does the distribution of amobarbital. Thus, the most accurate method of determining correlation of brain region with memory function during intracarotid amobarbital injection involves the use of a tracer such as HMPAO.