Ivy V. Estabrooke,1 Kristen Mordecai,2 Pauline Maki2 & Michael T. Ullman1
1 Georgetown University, 2 National Institute of Aging
We previously reported sex differences in the neurocognition of language. Converging evidence from psycholinguistic, neuropsychological and electrophysiological studies indicated that women tend to memorize at least certain complex linguistic representations (e.g., walked) in a temporal-lobe based lexical/declarative memory system, whereas men generally compute them compositionally (walk + -ed) in real-time with a frontal/basal-ganglia grammatical/procedural system (1, 2).
We predicted these findings on the basis of a previously-demonstrated female advantage at remembering words (3). This advantage depends on declarative memory, especially the hippocampus (4, 5). Word-learning and hippocampal sex differences are linked to estrogen. Estrogen improves word memory in women (6, 7) and men (8), and strengthens hemodynamic, physiological and biochemical correlates of hippocampal learning (9, 10). Testosterone, which is the main source of estrogen in men, improves their word memory (11).
Estrogen levels in males are lower than in females, prior to menopause (3). Thus our findings could be explained by organizational sex differences in utero or by differences in circulating estrogen levels during childhood, resulting in a greater tendency of girls than boys to memorize complex forms. Additionally, perhaps both sexes memorize complex forms, but females have superior lexical retrieval abilities, possibly due to higher levels of estrogen.
To examine these issues, we carried out a randomized, double-blind, placebo-controlled study in which 12 elderly men and 10 age-matched post-menopausal women were given hormone replacement therapy (women: estrogen; men: testosterone) and placebo, each for three months. They were tested on past-tense production after each three-month period.
Hormone therapy in both sexes led to an increase in the production rate of real and novel irregular forms (break-broke, spling-splung), p<.05, with no interaction between treatment (placebo/hormone) and sex (male/female). This suggests that estrogen in women and testosterone in men improve memory retrieval or processing. The absence of a treatment-by-sex interaction indicates that the two hormones have similar effects, suggesting the same mechanism --- presumably estrogen-modulated --- in the two cases. Performance at novel regulars (plagged) decreased with hormone therapy (p<.05; no interaction with sex), likely due to increased irregularizations (plag-plog). Thus the sex hormones do not ameliorate grammatical/procedural functions. Importantly, accuracy at real regulars (swayed) increased in women but decreased in men as a result of hormone therapy (treatment-by-sex interaction, p=.057) --- supporting the view that women but not men memorize complex forms. In summary, estrogen in women and testosterone in men improve lexical retrieval/processing but not rule-based computations.
References
[1] M. T. Ullman (2001). Nature Reviews Neuroscience, 2, 717-726.
[2] M. T. Ullman et al., (submitted).
[3] D. Kimura (1999). Sex and Cognition. The MIT Press, Cambridge.
[4] L. R. Squire, B. J. Knowlton (2000). In The New Cognitive Neurosciences, M. S. Gazzaniga (Ed.), pp. 765-780. The MIT Press, Cambridge, MA.
[5] T. Pfluger et al. (1999). Epilepsia, 40, 414-423 (1999).
[6] P. M. Maki, A. B. Zonderman, S. M. Resnick (2001). American Journal of Psychiatry, 158, 227-233.
[7] B. B. Sherwin (1996). Proceedings of the Society for Experimental Biology and Medicine, 217, 17-22.
[8] C. Miles, R. Green, G. Sanders, M. Hines (1998). Hormones and Behavior, 34, 199-208.
[9] C. S. Woolley, P. A. Schwartzkroin (1998). Epilepsia, 39, S2-8.
[10] P. M. Maki, S. M. Resnick (2000). Neurobiology of Aging, 21, 373-383.
[11] M. M. Cherrier et al. (2001). Neurology, 57, 80-88.