Comparative mapping of higher visual areas in monkeys and humansOutline placeholderComplications in relating human f&?h=1pt;MRI to monkey studiesMonkey f&?h=1pt;MRI fills a missing linkStrategies for comparing human and monkey visual cortical systemsDefining cortical areasCriteria for inferring homologySurface-based approaches to homology evaluationConserved early visual areasA mixed bag: the mid-level visual areasLikely homology: area V3AhMT+, a complex in need of subdivisionIn search of the human homologue of macaque V4Functional differences in higher-order regionsThe IT complex: an example of ‘regional’ homologyThe IPS region: can the discrepancies be resolved?ConclusionsAcknowledgementsReferencesComparative mapping of higher visualareas in monkeys and humansGuy A. Orban1, David Van Essen2and Wim Vanduffel1,31Lab Neuro- en Psychofysiologie, K.U. Leuven, Medical School, Campus Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium2Department of Anatomy and Neurobiology, Washington University, Medical School, St Louis, Missouri 63110, USA3MGH/MIT/HMS Athinoula A. Martino’s Center for Biomedical Imaging, 13th Street, Bldg 149, Charlestown,Massachusetts 02129, USAThe advent of functional magnetic resonance imaging(fMRI) in non-human primates has facilitated compari-son of the neurobiology of cognitive functions in humansand macaque monkeys, the most intensively studiedanimal model for higher brain functions. Most of thesecomparative studies have been performed in the visualsystem. The early visual areas V1, V2 and V3, as well asthe motion area MT are conserved in humans. Beyondthese areas, differences between human and monkeyfunctional organization are increasingly evident. At theregional level, the monkey inferotemporal and intra-parietal complexes appear to be conserved in humans,but there are profound functional differences in theintraparietal cortex suggesting that not all its constitu-ent areas are homologous. In the long term, fMRI offersopportunities to compare the functional anatomy of avariety of cognitive functions in the two species.The advent of functional imaging, initially PET and nowmainly fMRI has greatly enhanced our ability to explore theneurobiological basis of cognitive function in humans [1].However, fMRIonly indirectlyreflects neuronal activity andhas limited spatial and temporal resolution. Hence, theinterpretation of human fMRI data frequently draws on thevast knowledge obtained over recent decades by invasivebrain studies in non-human primates, especially themacaque. The study of the visual system serves as a modelin this respect, for a combination of reasons. Extensivepsychophysical studies have shown that many aspects ofvisual perception are remarkably similar in the two species.The visual cortex is heavily developed in primates: approxi-mately 50% of cerebral cortex in macaque and 20–30% inhumans is devoted to vision, compared with about 3% forauditionin monkeys and8% in humans [2]. The visual cortexofmacaquemonkeyshasbeeninvestigated intensively, morethan any other cerebral system. This has generated aplethoraoffunctional parcellations ofmacaquevisualcortex:30 or more anatomically and/or functionally distinct areashave been described (for a review see [2]). Finally, vision isfrequently used to study cognitive processes such asdiscrimination [3], attention [4], working memory [5] anddecision processes [6].Complications in relating human fMRI to monkey studiesEven in this favorable case of the visual system, establish-ing the relationship between non-invasive functionalimaging in humans and invasive single-cell, lesion oranatomical studies in monkeys is far from straight-forward. Making comparisons across species and tech-niques raises several challenges (see Box 1). Humans andmacaques diverged from a small-brained common ancestor, 30 million years ago [7]. Because the ensuing expansionof cerebral cortex was far greater in the human lineage,the cortex is ten-fold greater in surface area and also farmore convoluted in humans compared with macaques. Thedifferences are not simply a matter of scale, but instead arelikely to involve divergences in the number of visual areasand in how they are functionally specialized.In single-cell studies, inferences about the function(s)of a visual area are often based on tuning curves or,more generally, selectivity for various stimulus dimen-sions. There can be considerable diversity in the typesand degree of selectivity encountered in the neuronalBox 1. How close can one get in comparing human andmonkey using fMRI?Even when using fMRI in awake subjects, the experimental pro-cedures for the two species differ in several respects. Typically,monkeys sit in a sphinx position viewing a projection screen directly[15], whereas humans lie on their back viewing the screen through amirror. There are also differences in head immobilization and reward.The coils used to measure the MR signals are different, as arethe signals themselves. The use of a contrast agent (MION) in themonkey enhances sensitivity and signal localization [15,77] com-pared with the blood oxygen-level dependent effect (BOLD) used inhumans. To compensate for these latter differences, one can esti-mate a scale factor for the sensitivity by comparing MR signals in alandmark region such as V1 [21].The number of subjects is usually higher in human studies, but thenumber of functional volumes sampled in each subject is smaller.The standard spatial resolution of the fMRI measure is lower inhumans than in monkeys, but with high-field scanners humans canbe scanned with the same resolution (2 £ 2 £ 2 mm) as monkeys [20].There are important differences in how functional data are registeredto the anatomical MRI, in the details of the statistical analysis, andhow data are registered across individuals. Monkeys might payrelatively less attention to the stimuli than humans because only thefixation point controls their behavior. Therefore some experimentshave been repeated with monkey and human subjects performing atthe same level in a very demanding high acuity task [14,20].Corresponding author: Guy A. Orban ([email protected]).Available online 11 June 2004Review TRENDS in Cognitive Sciences Vol.8 No.7 July 2004www.sciencedirect.com 1364-6613/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2004.05.009population within a given area [8]. Functional MRI pro-vides an indirect measure of spike plus synaptic activity [9]of