Criterion-based validity of an intracarotid amobarbital recognition-memory protocol

We tested whether the behavioral components of an Intracarotid Amobarbital Protocol (IAP) had criterion validity. It was hypothesized that a recognition-memory test designed for intracarotid injections and used to predict the risk of global amnesia before an elective temporal lobectomy should also identify persons who are severely amnesic due to other neurologic causes. Divergent validity predicts that speech tasks would be unaffected by amnesia. Test-retest reliability also was measured.

Fifteen persons with severe amnesia were administered four alternate forms of a yes/no recognition-memory protocol and a speech protocol. No drug injection occurred. Standardized neuropsychological tests were used to divide the amnesic group into those with Global Amnesia (i.e., retain no ongoing memories), Severe Amnesia (i.e., memory impaired), and Amnesia Plus (severe amnesia plus other neuropsychologic deficits).

Two persons with Global Amnesia obtained scores at or below chance (i.e., failed) on the memory protocol. Unexpectedly, 12 of 13 severely amnesic persons obtained near-perfect memory scores. Amnesia had no impact on the speech protocol. Pass/Fail outcomes were highly correlated across all four sets.

A four-item IAP memory protocol showed good reliability and criterion validity in identifying the rare condition of Global Amnesia, but it was insensitive to other disabling, severe amnesic disorders. This IAP memory protocol might have validity in predicting a postsurgical Global Amnesic disorder, but it did not identify and therefore could not predict other more common severe amnesic disorders.