Dissociating aspects of verbal working memory within the human frontal lobe: Further evidence for a “process-specific” model of lateral frontal organization

Psychobiology 2000.28 (2).146-155 Dissociating aspects of verbal working memory within the human frontal lobe: Further evidence for a "process-specific" model of lateral frontal organization ADRIAN M. OWEN,ANDY C. H. LEE, and EMMAJ. WILIJAMS MRC Cognition and Brain Sciences Unit, Cambridge, England and University of Cambridge, Cambridge, England There is now converging evidence that suggests that working memory processes within the dorsolateral and ventrolateral frontal cortices are organized according to the type of processing required, rather than according to the nature (i.e., domain) of the information being processed, as has been widely assumed. For example, recent positron emission tomography (PET) studies have demonstrated that either, or both, of these two lateral frontal areas can be activated in spatial working memory tasks, depending on the precise executive processes that are called upon by the task being performed. Moreover, in a recent study using functional magnetic resonance imaging, performances of visual spatial and visual nonspatial working memory tasks were shown to involve identical regions of the lateral prefrontal cortex when all the factors unrelated to the type of stimulus domain were appropriately controlled. These results concur fully with recent reviews of the imaging literature, which have demonstrated that spatial and nonspatial working memory studies, in general, have produced a widely distributed pattern of overlapping activation foci within these lateral frontal regions. In this study, the effects of varying the executive requirements of a simple verbal working memory task (forward vs. backward digit span) were explored in 8 subjects, using PET, in order to establish whether this model generalizes to the verbal domain. As was expected, during forward digit span, significant activation was observed within the midventrolateral frontal cortex, but not within the middorsolateral frontal cortex. In contrast, during backward digit span, significant activation was observed in both regions. The results provide further evidence that the middorsolateral and midventrolateral frontal cortical areas make distinct functional contributions to memory and that this corresponds, in psychological terms, to a fractionation of working memory processes. In recent years, the term working memory has come to comparison between species is further compounded by be used in various ways: to describe a cognitive system Honig's (1978) definition of working memory, as applied for both the temporary storage and the manipulation of invariably in rat studies using Olton's radial arm maze remembered information (e.g., Baddeley, 1986), to de- (Olton, 1982), which also emphasizes the organizational scribe the type of memory that is active and relevant only or executive component of task performance. Neverthefor a short period of time (e.g., Fuster, 1995; Goldman- less, in the absence of a precise definition, few disagree Rakic, 1995), and most specifically, to describe the pro- that working memory is a fundamental set of processes cess by which a remembered stimulus is held on line to and an integral component of many cognitive operations, guide behavior in the absence of external cues or prompts -trom complex decision making to selective attention (Bad(Goldman-Rakic, 1987). In part, this descriptive vari- deley, 1986). Although evidence from the study of patients with exability reflects the relative interests of those working with different primate species: Psychologists working mainly cisions of the frontal cortex, from lesion and electrowith humans often emphasize the organizational or higher physiological recording work in nonhuman primates (for order aspects of working memory tasks, whereas those reviews, see Goldman-Rakic, 1994, and Petrides, 1994), working with nonhuman primates tend to focus on those and from functional neuroimaging studies in humans aspects of task performance related to the on-line retention suggests that the lateral frontal cortex plays a critical role or short-term storage of information. The problem of in certain aspects of working memory, no consensus has been reached regarding the fractionation of functions within this region (e.g., Cohen et aI., 1994; Cohen et aI., A.C.H.L. is in the Department of Experimental Psychology, Cam1997; Courtney, Ungerleider, Keil, & Haxby, 1996, 1997; bridge, and E.J.W. is at the Wolfson Brain Imaging Centre, Cambridge. Funahashi, Bruce, & Goldman-Rakic, 1989, 1990; Gold, Correspondence concerning this article should be addressed to A. M. Berman, Randolph, Goldberg, & Weinberger, 1996; GoldOwen, MRC Cognition and Brain Sciences Unit, 15 Chaucer Road, berg, Berman, Randolph, Gold, & Weinberger, 1996; Cambridge, CB2 2EF, England (e-mail: . ac.uk). Jonides et aI., 1993; McCarthyetal., 1994; Owen, Downes, Copyright 2000 Psychonomic Society, Inc. 146 WORKING MEMORY IN HUMAN LATERAL FRONTAL CORTEX Sahakian, Polkey, & Robbins, 1990; Owen, Doyon, Petrides, & Evans, 1996; Owen, Evans, & Petrides, 1996; Owen, Morris, Sahakian, Polkey, & Robbins, 1996; Owen, Sahakian, Semple, Polkey, & Robbins, 1995; Petrides, Alivisatos, Evans, & Meyer, 1993; Petrides & Milner, 1982; Rao, Rainer, & Miller, 1997; Smith, Jonides, & Koeppe, 1996; Smith et aI., 1995; Stem et aI., 2000; Sweeney et aI., 1996; Wilson, Scalaidhe, & GoldmanRakic, 1993; for reviews, see Owen, 1997, Rushworth, Nixon, Eacott, & Passingham, 1997, and Rushworth & Owen, 1998). Until recently, one prevalent view has been that working memory processes are organized according to the type (e.g., domain) of information being pr()cessed, with dorsolateral frontal regions being concerned prin.:-cipally with memory for spatial material, whereas the anatomically and cytoarchitectonically distinct ventrolateral frontal regions subserve memory for nonspatial material (Goldman-Rakic, 1987, 1994, 1995; Levy & Goldman-Rakic, 1999; see also Courtney, Petit, Maisog, Ungerleider, & Haxby, 1998; Courtney et al., 1996, 1997). In general, however, both functional neuroimaging studies in humans (for reviews, see D'Esposito, Aguirre, et aI., 1998; Owen, 1997, in press, and Rushworth et aI., 1997; see also Rushworth & Owen, 1998) and recent electrophysiological recording studies in the monkey (Rao et aI., 1997) have failed to support this domain-specific model oflateral frontal lobe function. For example, in one recent review of the literature (Owen, 1997), the stereotaxic coordinates of activation foci reported within the dorsolateral and ventrolateral frontal cortices during 14 spatial working memory tasks (Baker, Frith, Frackowiak, & Dolan, 1996; Goldberg et aI., 1996; Jonides et aI., 1993; McCarthy et aI., 1994; McCarthy et aI., 1996; Owen, Evans, & Petrides, 1996; Smith et aI., 1996; Smith et aI., 1995) were compared with those observed during 6 nonspatial (visual) working memory tasks (Baker et aI., 1996; Courtney et aI., 1996; McCarthy et aI., 1996; Owen et aI., 1998; Petrides (...truncated)