Abstracts

Rationale: It is well established that the medial temporal region plays a primary role in episodic memory. Accordingly, there is a strong correlation between the integrity of the hippocampus and verbal memory performance in left (dominant) TLE patients. Less clear, however, is the relation between the hippocampus and naming. Although the hippocampus is not typically considered part of the language system, studies involving MR volumetrics and MR spectroscopy have shown a correlation between hippocampal integrity and visual object naming performance. We speculated that perhaps, this relation might reflect non-naming aspects of visual object naming. We sought to determine whether other factors, possibly more feasibly be related to hippocampal integrity (i.e., age of epilepsy onset, vocabulary knowledge), might underlie an apparent relation between naming and hippocampal integrity. Methods: Subjects were 149 left (dominant) temporal lobe epilepsy patients (83 with Medial Temporal Sclerosis ( MTS ), mean age 38.3, (SD= 12.0), mean education 14.7 (SD= 2.6), mean Full Scale IQ 97.4 (SD=14.4) who underwent neuropsychological evaluation at the Columbia Comprehensive Epilepsy Center. Naming was assessed using the Visual Naming Test, the Auditory Naming Test and the Boston Naming Test. Naming scores were log transformed as raw scores were not normally distributed. T tests assessed group (MTS vs. nonMTS) differences in naming and vocabulary performance. Pearson correlations followed by multivariable linear regression analyses assessed the contribution of MTS, epilepsy onset age, years of education, and vocabulary knowledge to naming performance. Results: Naming performances were higher among patients without MTS relative to those with MTS for the ANT (P = .002) and BNT (P= .006), but not the VNT (P =.12). Naming performances correlated with Onset Age (ANT and VNT: r = -.19, P= .02, BNT r = .22, P= .01), Vocabulary (ANT: r = -.51, P=.001, VNT: r = -.34, P= .001, BNT( r = .67, P= .001), and Education (ANT: r = -.51, P=.001, VNT: r = -.34, P= .001, BNT r = .67, P= .001), but not with chronological age. Although simple regression analyses for all naming measures suggested a significant influence of MTS, presence/absence of MTS no longer accounted for a significant portion of the variance when Vocabulary and/or Education were added to the model.Conclusions: Although the presence/absence of MTS appeared to influence naming performance, further analyses suggested that other factors might mediate this apparent relation. That is, naming performance might not be mediated directly by the hippocampus; rather vocabulary knowledge (i.e., verbal learning) might be more directly associated with hippocampal function, which in turn, is related to naming performance.