Functional topography of a distributed neural system for spatial and nonspatial information maintenance in working memory
Introduction
Working memory (WM) is the means by which a limited amount of information is actively maintained and processed for a short period of time. Single cell recordings of monkey prefrontal cortex during delayed match to sample and delayed response tasks have shown that cells exhibit stimulus specific sustained activity over delay periods, presumably reflecting information maintenance (e.g. [20], [22], [23], [43], for reviews see [24], [25]). Functional magnetic resonance imaging (fMRI) studies of human subjects also show sustained activity over delay periods of WM tasks in prefrontal cortex (e.g. [9], [11], [78]).
The role of regions within the prefrontal cortex in maintaining different types of information in WM is under debate, however. While there appear to be dissociations between prefrontal regions based on the type of processing required, independent of the type of information involved [11], [14], [49], [56], [70], the question remains whether those prefrontal areas involved in simple maintenance of information also have a functional organization according to the type of information maintained. It has been proposed that the visual system is divided into two (interconnected) processing streams. The ventral pathway, including occipitotemporal cortex, is thought to represent perceptual properties that are important for object identification, such as color, texture, and shape. The dorsal pathway, including parietal cortex, on the other hand, processes information regarding the spatial locations of objects, the spatial relationships among objects, and the guidance of motor movements toward objects [26], [33], [73], [74]. One model proposes that this ventral/dorsal, object/spatial dissociation extends to the prefrontal cortex, resulting in separate functionally specialized areas for object and spatial WM maintenance (for review see [38]). Single neurons in the monkey dorsolateral prefrontal cortex, near the posterior end of the principal sulcus, have been shown to exhibit spatially selective delay-related activity during visuospatial delayed response tasks (e.g. [20], [23], [37]). In contrast, neurons in the inferior convexity in the ventrolateral prefrontal cortex that do not respond during spatial delayed response tasks, do often respond selectively to textures, shapes, and faces, and respond best to stimuli presented foveally. Neurons of this sort are very seldom found in the principal sulcus. These neurons often have sustained stimulus-selective activity in object–response association tasks and even in tasks that do not have a required memory or response component ([47], [77], review in [38]).
Behavioral deficits resulting from lesions in both humans and monkeys also suggest a different neuroanatomical organization for spatial and nonspatial WM. In monkeys, lesions of cortex within the posterior portion of the principal sulcus impair spatial WM but not object WM [21], [41], [52], [53]. Results from lesions of ventral prefrontal cortex (the inferior convexity) can impair performance on both spatial and object WM tasks, although the nature and duration of the deficits is unclear [38], [42], [66]. In humans, a recent report described a patient with a lesion of the right superior prefrontal cortex that showed a selective deficit for spatial WM. This patient had no impairment of verbal or object WM and no visual perceptual or attention deficit [7]. A different individual patient with a lesion restricted to inferior lateral prefrontal cortex who had a selective deficit for nonspatial WM has also been reported [5]. Transient “lesions” induced by repetitive transcranial magnetic stimulation in humans demonstrated a double dissociation, with dorsal medial stimulation (near the superior frontal sulcus (SFS)) producing selective impairment of a spatial WM task and ventral lateral stimulation producing selective impairment of a face WM task [45].
Accordingly, many imaging studies in humans have shown that the ventral prefrontal cortex, such as the inferior and middle frontal gyri (IFG/MFG), is most consistently activated during WM for faces [11], objects other than faces [6], [39], [67], and verbal information [3], [17], [35], [54], [61], [63], [68]. More dorsal prefrontal cortex, specifically an area of the superior frontal sulcus, has been most consistently activated during spatial WM tasks [4], [8], [10], [34], [40], [46], [50], [55], [68], [78]. Significant double dissociations in the amount of activation in dorsal and ventral prefrontal cortex for spatial and nonspatial information have been shown with both visual [10], [12] and auditory [2] stimuli.
It should be noted, however, that both dorsal and ventral prefrontal regions are frequently activated for both spatial and object WM tasks relative to low-level control tasks and several imaging studies in humans have found no significant differences in the patterns of activation for these two types of tasks [4], [6], [39], [46], [50], [59], [67]. Single cell recording data also suggest that many cells across both dorsal and ventral prefrontal cortex maintain both spatial and object information and it has been suggested that object and spatial information are integrated, rather than segregated into specialized areas, in prefrontal cortex. Individual cells have been found in the prefrontal cortex that show sample-specific activity over the delay period for either object, location, or the combination of object and location information with no obvious topological organization regarding the anatomical distribution of these cell populations [44], [62], [64].
The reason for these apparently conflicting results regarding the functional segregation or the integration of WM maintenance in the prefrontal cortex is unclear. Human neuroimaging studies that did not find significant differences in the activation between object and spatial WM tasks within the prefrontal cortex used either geometric shapes or patterns as stimuli in the object task [46], [51], [58], [59]. A study by Courtney et al. [12] which demonstrated a clear dorsal–ventral double dissociation for spatial and object identity information used faces as stimuli. In the monkey frontal cortex, the representation of faces in particular appears to be restricted to the ventral inferior convexity [47], [48]. If the representation of faces is restricted to the ventral prefrontal cortex while the other object representations are not, then the dorsal–ventral double dissociation between face and spatial WM observed in the Courtney et al. [12] study could be specific to faces and not a general principle extending to all objects [60].
Some categories of objects may require a greater representation of spatial information than other categories of objects. If the objects to be remembered varied between sample and test in terms of spatial aspects (e.g. spatial relationships among parts), rather than in terms of nonspatial aspects (e.g. color or texture), this could lead to a different pattern of brain activation and different behavioral performance on tasks involving these different categories of objects. For example, greater activation of parietal cortex for houses than for faces has been observed previously in studies of object perception [28], [29], [31], [32]. This activation of parietal areas by house perception could reflect a distributed representation of houses across both ventral (nonspatial) and dorsal (spatial) visual areas. Alternatively, it could reflect an activation that is due to some spatial aspect of house stimuli (e.g. spatial navigation associations, spatial scale, or spatial variation between exemplars) that would not be necessary to the task of identifying the house. The distributed representation account would imply shared neural resources for a spatial location task and a house identity task while the incidental activation account would imply independent neural resources.
In the current study, we investigated the pattern of activation during WM tasks for face identity, house identity, and spatial location of faces and houses using fMRI (Experiments 1–3). We also performed a behavioral experiment utilizing a dual task paradigm to determine the extent to which spatial location information can be maintained concurrently with separate face or house identity information (Experiment 4). If faces and houses place differential demands on the neural substrates necessary for location WM, then there should be different behavioral effects when attempting to maintain location information while simultaneously maintaining face or house information. The fMRI and behavioral results together suggest that the mechanisms for the maintenance of house identity information (and perhaps other objects) are distributed and overlapping with those that maintain spatial location information, while the mechanisms for maintenance of face identity information are relatively more independent. There is, however, a consistent functional topography that results in superior prefrontal cortex producing the greatest response during spatial WM tasks, and middle and inferior prefrontal cortices producing their greatest responses during object WM tasks, independent of the object type.
Section snippets
Subjects
Subjects for the preliminary behavioral testing of the tasks used in the fMRI experiments (Experiment 1 and 2, N=11; Experiment 3, N=7), and for the behavioral dual-task experiment (N=10), consisted of Johns Hopkins University students. Participants were recruited from a pool of students volunteering to do psychology experimentation in return for extra credit in undergraduate psychology classes.
Subjects for the fMRI portions of this study were non-smokers in good health that had no history of
Performance data for fMRI tasks
Prior to collecting fMRI data, all tasks were tested behaviorally on separate groups of subjects from those participating in the fMRI portion of the study. Technical difficulties during Experiment 1 prohibited the collection of behavioral data during scanning. Mean percent correct and reaction time for each of the tasks (including training performance for subjects in Experiment (1) is given in Table 1. There were no significant differences in the accuracy for stimulus type (faces versus
Discussion
It is evident from all the three fMRI experiments that a large-scale, distributed cortical network underlies the active maintenance (storage and rehearsal) of visual information in WM, as reported in previous studies. Areas in temporal, parietal, and both ventral and dorsal prefrontal cortices showed greater activity for either the spatial or object WM tasks relative to the sensorimotor control task. However, there exists a differential pattern of response associated with the maintenance of
Acknowledgements
The authors wish to thank the entire staff of the F.M. Kirby Research Center for Functional Brain Imaging for assistance with data collection and storage. We thank Dr. Ed Awh for suggestions concerning behavioral testing, Dr. Steven Yantis and John Serences for assistance with the collection of behavioral data, and Dr. Scott Slotnick and John Serences for programming assistance. We thank Drs. Ed Awh and James Haxby for helpful comments on an earlier draft of the manuscript. This work was
References (78)
- et al.
An area within human ventral cortex sensitive to “building” stimuli
Neuron
(1998) - et al.
Remembering what but not where: independence of spatial and visual working memory in the human brain
Cortex
(2001) AFNI: software for analysis and visualization of functional magnetic resonance neuroimages
Computers and Biomedical Research
(1996)- et al.
Maintenance versus manipulation of information held in working memory: an event-related fMRI study
Brain and Cognition
(1999) - et al.
The parahippocampal place area: recognition, navigation, or encoding?
Neuron
(1999) - et al.
Analysis of fMRI time-series revisited
NeuroImage
(1995) - et al.
How many subjects constitute a study?
NeuroImage
(1999) Cellular basis of working memory
Neuron
(1995)- et al.
Separate visual pathways for perception and action
Trends in Neuroscience
(1992) - et al.
The effect of face inversion on activity in human neural systems for face and object perception
Neuron
(1999)
Non-spatial memory after selective prefrontal lesions in monkeys
Brain Research
A re-examination of the effects of frontal lesion on object alternation
Neuropsychologia
The prefrontal cortex: complex neural properties for complex behavior
Neuron
Working memory for letters, shapes, and locations: fMRI evidence against stimulus-based regional organization in human prefrontal cortex
NeuroImage
Delayed matching after selective prefrontal cortex lesions in monkeys (Macaca mulatta)
Brain Research
An fMRI investigation of cortical contributions to spatial and nonspatial visual working memory
NeuroImage
Activity in ventrolateral and mid-dorsolateral prefrontal cortex during nonspatial visual working memory processing: evidence from functional magnetic resonance imaging
NeuroImage
‘What’ and ‘where’ in the human brain
Current Opinion in Neurobiology
Temporal isolation of the neural correlates of spatial mnemonic processing with fMRI
Cognitive Brain Research
“What” and “where” in the human auditory system
Proceedings of the National Academy of Sciences of the United States of America
Dissociation of storage and rehearsal in verbal working memory: evidence from positron emission tomography
Psychological Science
Active representation of shape and spatial location in man
Cerebral Cortex
Dissociation of working memory from decision making within the human prefrontal cortex
Journal of Neuroscience
Dissociation of mnemonic and perceptual processes during spatial and nonspatial working memory using fMRI
Human Brain Mapping
Distribution of cortical activation during visuospatial n-back tasks as revealed by functional magnetic resonance imaging
Cortex
Temporal dynamics of brain activation during a working memory task
Nature
Object and spatial visual working memory activate separate neural systems in human cortex
Cerebral Cortex
Transient and Sustained activity in a distributed neural system for human working memory
Nature
An area specialized for spatial working memory in human frontal cortex
Science
A cortical representation of the local visual environment
Nature
A positron emission tomography study of the short-term maintenance of verbal information
Journal of Neuroscience
Mnemonic coding of visual space in the monkey’s dorsolateral prefrontal cortex
Journal of Neurophysiology
Dorsolateral prefrontal lesions and oculomotor delayed-response performance: evidence for mnemonic “scotomas”
Journal of Neuroscience
Neuron activity related to short-term memory
Science
Unit activity in prefrontal cortex during delayed-response performance: neuronal correlates of transient memory
Journal of Neurophysiology
Searching for a baseline: functional imaging and the resting human brain
Nature Reviews Neuroscience
Distributed and overlapping representations of faces and objects in ventral temporal cortex
Science
Cited by (146)
Neural representations of social valence bias economic interpersonal choices
2020, NeuropsychologiaCitation Excerpt :Interestingly, these areas were part of a large cluster that also showed increased activation during the maintenance of social information, as revealed by univariate results. In non-social paradigms, the lPFC has been related to working memory maintenance (Morgan et al., 2013; Sala et al., 2003) and other forms of cognitive control (e.g., Reverberi et al., 2012). The IFG specifically has also been associated with the selection of semantic information (Jefferies, 2013; Wagner et al., 2001), and it is also involved in the expectation to perform different non-social tasks employing verbal material (e.g., González-García et al., 2017; Sakai and Passingham, 2006).
The neural basis of complex audiovisual objects maintenances in working memory
2019, NeuropsychologiaDistinct effects of dopamine vs STN stimulation therapies in associative learning and retention in Parkinson disease
2016, Behavioural Brain ResearchWorking Memory, Neural Basis of
2015, International Encyclopedia of the Social & Behavioral Sciences: Second Edition