The role of prefrontal cortex in working-memory capacity, executive attention, and general fluid intelligence: An individual-differences perspective

RANDALL W. ENGLE ) 0 1 2 0 This work was supported by Grants F49620-93-1-0336and F49620- 97-1 from the Air Force Office of Scientific Research and Grant RO1- HD27490-01A1 from the National Institute of Child Health and Human Development. We are indebted to Alan Baddeley , Todd Braver, Andrew Conway, John Duncan, Timothy Salthouse, and Jeffrey Toth for their helpful comments on earlier versions of this manuscript. Correspon- dence concerning this article should be sent either to M. J. Kane, De- partment of Psychology , P. O. Box 26164, University of North Carolina at Greensboro , Greensboro, NC 27402-6164 ,or to R. W. Engle, School of Psychology,Georgia Institute of Technology , Atlanta, GA 30332-0170 ( 1 Georgia Institute of Technology , Atlanta, Georgia 2 MICHAEL J. KANE University of North Carolina , Greensboro, North Carolina We provide an executive-attention framework for organizing the cognitive neuroscience research on the constructs of working-memory capacity (WMC), general fluid intelligence, and prefrontal cortex (PFC) function. Rather than provide a novel theory of PFC function, we synthesize a wealth of singlecell, brain-imaging, and neuropsychological research through the lens of our theory of normal individual differences in WMC and attention control (Engle, Kane, & Tuholski, 1999; Engle, Tuholski, Laughlin, & Conway, 1999). Our critical review confirms the prevalent view that dorsolateral PFC circuitry is critical to executive-attention functions. Moreover, although the dorsolateral PFC is but one critical structure in a network of anterior and posterior attention control areas, it does have a unique executiveattention role in actively maintaining access to stimulus representations and goals in interference-rich contexts. Our review suggests the utility of an executive-attention framework for guiding future research on both PFC function and cognitive control. - The frontal lobes reach their phylogenetic and ontogenetic peak in adult Homo sapiens, where they occupy between 30% and 40% of the neocortical area (see, e.g., Brodmann, 1925; Damasio, 1991; Fuster, 1988; GoldmanRakic, 1987). Such evolutionary and physical prominence has led many theorists to assign the highest of cognitive capabilities, and even the highest qualities of humanity itself, to the frontal cortex (e.g., Goldstein, 1936, 1944; Halstead, 1947; Rylander, 1939). However, early clinical research on patients with frontal lobe damage indicated that such injury did not affect intelligence, at least as broadly defined by IQ test batteries (e.g., Ackerly, 1937; Hebb, 1939, 1945; Hebb & Penfield, 1940). Such null findings stand in stark contrast to the everyday cognitive difficulties reported by many patients with frontal damage, particularly by those with damage to the prefrontal cortex (PFC; see, e.g., J. M. Harlow, 1848; Lezak, 1983; Luria, 1966; Shallice & Burgess, 1991a, 1991b). Indeed, a more recent body of clinical observations and experimental research suggests that PFC injury and disease creates a formidable array of cognitive deficits. Such deficits include (but are not limited to) problems of attention, motor control, spatial orientation, short-term memory, temporal and source memory, metamemory, associative learning, creativity, perseveration, and reasoning (for reviews, see Fuster, 1988; Goldman-Rakic, 1987; A. C. Roberts, Robbins, & Weiskrantz, 1998; Stuss & Benson, 1984; Wise, Murray, & Gerfen, 1996). In the sections that follow, we will critically and comprehensively review evidence that general working-memory (WM) and executive-attentionfunctions are subserved by neural circuits centered in and passing through the PFC. There is broad agreement in the literature that PFC circuits, and perhaps dorsolateral prefrontal cortex (dPFC) cells in particular, are critical for WM functions. From our perspective, the role of dPFC in WM is to maintain information in a highly active, easily accessible state. This maintenance is particularly important in the presence of interference, and it may be crucial in blocking the effects of distraction. Moreover, as we have argued elsewhere (e.g., Engle, 2001, 2002; Engle, Tuholski, Laughlin, & Conway, 1999), we view WM capacity, or the capability for executive attention, as the psychological core of the statistical construct of general fluid intelligence, or psychometric Gf. In this review, then, we will also evaluate evidence suggesting the importance of the dPFC to general fluid ability. We will further speculate that, because dPFC is critical to WM capacity and to Gf, normal individual differences in WM capacity and in Gf may be mediated by normal individual differences in dPFC functioning (see also Engle, Kane, & Tuholski, 1999; Engle & Oransky, 1999). As the quotes that began this paper may suggest, we do not claim to present a novel theory of PFC function, but rather we present an organizing framework for reviewing prior research and suggesting fruitful avenues for future work. Indeed, the lens through which we examine the literature has much in common with several prominent views of dPFC function, particularly those of Baddeley (1996), Dempster (1991, 1992), Duncan (1993, 1995), Fuster (1988, 1996), Goldman-Rakic (1987), Malmo (1942), E. K. Miller & Cohen (2001), R. J. Roberts and Pennington (1996), Shallice and Burgess (1991b), Shimamura (2000), Smith and Jonides (1997), and Stuss, Shallice, Alexander, and Picton (1995). Our contribution here is novel, however, not just in the comprehensiveness of the review, but also in providing a unified perspective on four broad, overlapping constructs: WM capacity, attention control, fluid intelligence, and PFC function. Relations within only a subset of these constructs have been explored in detail beforefor example, among attention, intelligence, and PFC (Dempster, 1991, 1992; Duncan, 1993, 1995). Moreover, although aspects of WM function have long been linked to the PFC (see, e.g., Jacobsen, 1935, 1936), the individual-differences construct of WM capacity has not. To summarize our view, the WM construct is assumed to be a hierarchical system involving short-term-memory (STM) representational components plus a general, executive-attention component (see Baddeley & Hitch, 1974; Baddeley & Logie, 1999; Cowan, 1995, 1999). Span tasks that reflect WM capacity are thought to reflect the contributions of both STM and executiveattention components (see, e.g., Daneman & Carpenter, 1980; Turner & Engle, 1989). However, the covariation among WM-span tasks and tasks of higher order cognition reflects primarily the executive-attention component in the system, and less the STM components. Normal individual differences in WM span are widely found to correlate with many facets of higher order cognition, including language comprehension, reasoning, and Gf. We believe that these correlations are driven by individual differences in executive attention. Thus, when we use the term WM capacity, which we do for his (...truncated)