Regional cortical thickness matters in recall after months more than minutes

Abstract

The aim of this study was to determine the role of regional cortical thickness in recall of verbal material over an extended time period. MRI scans of healthy adults of varying ages were obtained. Two scans were averaged per person to achieve high spatial resolution, and a semi-automated method for continuous measurement of thickness across the entire cortical mantle was employed. Verbal memory tests assessing recall after 5 min, 30 min, and a mean interval of 83 days were administered. A general linear model (GLM) of the effects of thickness at each vertex on the different memory indices was computed, controlling for gender, age, IQ, and intracranial volume. These analyses were repeated with hippocampal volume as an additional variable to be controlled for, to assess to which extent effects of cortical thickness were independent of hippocampal size. Minute effects of cortical thickness were observed with regard to shorter time intervals (5 and 30 min). However, even when controlling for the effects of hippocampal volume, higher recall across months was associated with thicker cortex of distinct areas including parts of the gyrus rectus, the middle frontal gyrus, the parieto-occipital sulcus and the lingual gyrus of both hemispheres. In addition, hemisphere-specific associations were found in parts of the right temporal and parietal lobe as well as parts of the left precuneus. This supports a unique and critical role of the thickness of distinct cortical areas in recall after months, more than after minutes.

Introduction

New magnetic resonance (MR) imaging and neuroanatomical quantification techniques has lead to much recent interest in relating volumetric brain characteristics to human abilities. The importance of volume of different brain areas for functional decline or maintenance in normal versus degenerative conditions has now been well documented (see, e.g., Jack et al., 2004, Fischl et al., 2002). There seems to be less consistent data, however, on the importance of normal individual differences in cerebral volumetric characteristics in healthy persons. Largely, two broad classes of studies have emerged: One focuses on the role of general brain volume or cortical volume in general ability (Willerman et al., 1991, Andreasen et al., 1993, Raz et al., 1993, Harvey et al., 1994, Wickett et al., 2000, Egan et al., 1994, Egan et al., 1995, Flashman et al., 1998, Thompson et al., 2001, MacLullich et al., 2002, Posthuma et al., 2002, Walhovd et al., 2005). The other class of studies mainly focuses on the role of specific brain structures in prediction of specific cognitive abilities. The relationship between frontal lobe volume and neuropsychologically defined frontal lobe function has been extensively studied (see van Petten et al., 2004, for an overview of 11 such studies), as well as the role of the hippocampal volume in memory performance (e.g., Raz et al., 1998, de Toledo-Morrell et al., 2000, Hackert et al., 2002, Golomb et al., 1994, Golomb et al., 1996, Rosen et al., 2003, Tisserand et al., 2000, Torres et al., 1997).

While the first class of studies, relating gross brain volumes to general ability, have lent support to a brain size–behavior relationship, robust relationships between specific cognitive abilities and morphometric characteristics have not been consistently established. For instance, the importance of hippocampus for mnemonic function is well documented (e.g., Scoville and Milner, 1957), but there is no support for a simple “bigger is better” perspective for the relationship between normal individual differences in hippocampal size and memory scores at retention intervals of an hour or less (see van Petten, 2004, for a review). Further, van Petten et al. (2004) did not find positive relationships between memory performance and several gyri of known importance for memory function in the frontal and temporal parts of neocortex. However, theoretical and empirical accounts imply that memories are strengthened and maintained across an extended time period by hippocampal–cortical interactions (Buzsaki, 1996, Kali and Dayan, 2004). Consolidation takes place over several days (Riedel and Micheau, 2001) or years (Haist et al., 2001), and while hippocampus may be involved over a prolonged interval (Buzsaki, 1996, Kali and Dayan, 2004, Ryan et al., 2001), theoretical accounts also imply that memories become increasingly cortically distributed with time. In a recent study, we found evidence for increased importance of hippocampal size across an extended retention interval with a mean of 11 weeks (Walhovd et al., 2005). However, evidence was not obtained for higher importance of cortical volume in memory across weeks independently of age, but only total cortical volume was included in that study. This poses the question whether volumetric characteristics of specific cortical areas might still be of importance in recall at such extended retention intervals. A relationship between thickness of specific cortical areas and retention may manifest itself more strongly across weeks, since some cortical areas likely are increasingly important in storage of memories with time (Buzsaki, 1996, Kali and Dayan, 2004). Further, as argued by Buckner and Wheeler (2001), it is important to determine whether cortical correlates of retrieval are dependent on characteristics of medial temporal lobe structures or not. Thus, the question whether regional cortical thickness variations is related to recall at different retention intervals was investigated also when the effects of hippocampal volume were controlled for.