Serial digit learning: Normative and clinical aspects

Abstract

The performances of 500 normal subjects on serial digit learning (8 or 9 digits over 12 trials) were obtained and normative standards in relation to age, educational level, and sex were established. The performances of 112 hospital control patients on digit span were also obtained to develop parallel empirical standards. The performances of 100 patients with hemispheric brain disease on these two tasks were then compared. Classificatory agreement (pass or fail) between the two tests was moderately high (69%), but serial digit learning was the more sensitive index of the presence of brain disease. The advantages of including digit learning in the mental status examination are discussed.