Viewing Cognitive Sex Differences as Computational Biases in Dorsal-Ventral Streams

Robert F. McGivern (SDSU) and Jaime A. Pineda (UCSD)

Until the latter part of the 20th century, reports of cognitive sex differences were generally viewed as reflections of socio-cultural learning, rather than as differences in the brain organization of men and women. Maccoby and Jacklin1 challenged this assumption in 1974 in their review of more than a thousand studies conducted in the preceding decade that reported results for men and women in cognitive, social, and emotional behavior. In the cognitive realm, two consistent sex differences emerged: women generally performed better than men on several types of verbal tasks, while men were more likely to excel on spatial tasks. Little evidence was found for sex differences in general cognitive abilities such as abstract reasoning, language comprehension, or planning. These findings, coupled with subsequent reports in animals and humans of anatomical sex differences, rapidly led to broad acceptance of a strong biological component to cognitive sex differences2.

Over the years, hundreds of additional studies have provided a clearer delineation of the sub-domains exhibiting reliable sex differences3. The female advantage in verbal skills is linked to processes such as episodic memory, verbal recall, object recognition memory, and communication skills, but not to others such as writing, general fluency, or language and reading comprehension. The male advantage in spatial skills centers on space relation tasks, such as mental rotation, targeting, and embedded figures, but not to object location memory, a ‘spatial’ skill that favors females. Overall, the cognitive skills that exhibit medium to large size effects are those involving space relations, object awareness, and social awareness, with female advantages found in the last two categories3.  However, variability on these types of tasks is significantly greater within than between the sexes, reinforcing the point that the term ‘sex difference’ is quite relative when applied to cognition.

The source of cognitive sex differences has been broadly debated.  Cognitive neuroscience, neurological, evolutionary, and sociobiological explanations are well-supported, but none adequately account for conceptually similar cognitive sex differences observed in animals. In other species, males show consistently better performance in spatial tasks than females, while females exhibit better object recognition skills. The same basic grouping is observed in human cognitive sex differences; space relation tasks favor males, while object recognition tasks favor females. Cross-species comparisons appear to be warranted based on the fact that the advantage in verbal skills for women falls into areas that reflect better encoding and retrieval of object characteristics such as color, facial expression, or even object location. Men’s advantage in space relations reflects better processing of real or abstract motion, as reflected in targeting and mental rotation tasks, respectively.  Viewed from this basic perceptual processing perspective, the cross-species cognitive similarities suggest that cognitive sex differences may arise from computational biases related to dorsal and ventral stream processing. 

Based on functional neurophysiology studies in the monkey, Ungerleider and Mishkin4 characterized two cortical pathways that serve the perception of ‘where’ and ‘what’ in the visual perception of the environment. The dorsal stream pathway (‘where’) extends from the primary visual cortex into the parietal lobes, providing information about the relational information of objects in the environment. The ventral stream (‘what’) extends from the primary visual cortex into the temporal lobes and is essential for the perception of individual objects and their associated characteristics (e.g., texture, color, shape, etc.). Both pathways operate interactively and in parallel, and project to the frontal cortex5. Parallel processing of ‘what’ and ‘where’ is also found in the auditory system6, suggesting that this early stage separation in perceptual processing is a relatively ubiquitous organizational principle in the neural substrates serving cognition.

The ‘what’ and ‘where’ formulation was later expanded by Goodale and Milner7 based on the results of patients with selective lesions within each functional processing stream. Patients with damage to the inferotemporal region can exhibit a severe visual agnosia, leaving them functionally blind because of an impaired ability to perceive color, form, shape or size of an object. However, these patients are able to navigate the environment unconsciously based on visual information sent to other regions.  In contrast, selective parietal lesions induce optic ataxia, while leaving the patient with full awareness of object characteristics in their field of vision. On this basis, Goodale and Milner reformulated dorsal and ventral stream as ‘vision for action’ and ‘vision for perception’, respectively. This view emphasizes the basic functional role each system serves in higher cognition. The ventral stream capacity to process objects and their characteristics provides the foundation for visual perception, whereas dorsal stream processing serves the higher order control of action, which may operate under conscious or unconscious control. 

Men and women may rely on differential computational strategies to process information in the environment about objects and object relationships, with women better primed to consciously perceive objects and their characteristics, and men better primed to process relational information among objects. Such a bias would imply greater unconscious processing of static object characteristics by women and greater unconscious processing of movement by men. Tasks that require processing against the computational bias of an individual would theoretically require greater conscious effort, which should be reflected in increased frontal activation. This is what was observed in a recent fMRI study of neural processing during mental rotations in normal men and women. Greater dorsal stream activation was found in men, contrasted with greater frontal activation in women8. The pattern was interpreted as indicating that men are more likely to rely on a bottom-up strategy and women on a top-down strategy, based on the assumption that frontal activation reflects the degree of conscious effort involved in the decision. From a behavioral perspective, differential strategies for male and females are also observed in solving navigation problems.  For both humans and animals, males are more likely to use cardinal relationships to navigate and environment whereas females are more likely to use landmarks9.

This pattern suggests that cognitive sex differences may stem from basic biases in individual computational styles, one of which emphasizes object characteristics and the other movement, as the first order of cognition. These biases may also be reflected in processing within specific frontal systems, including the mirror neuron system (MNS), which is active during the performance of actions, as well as their observation10. The MNS is further implicated in action comprehension, imitation learning, the ability to interpret facial expressions, language development, as well as empathy – behaviors that may play critical roles in emphasizing individual cognitive differences11.

Developmentally, the perceptual processing within the MNS and dorsal/ventral streams help form a sensorimotor foundation for higher cognitive processes. A bias in the development of these systems would be expected to have a rippling effect on the development of the higher cognitive processes that they serve. Gonadal steroids are the most likely biological influence to induce such a bias. Surges of testosterone in males occur during early prenatal development and at birth, acting on estrogen and androgen receptors within the cortex12. Animal and human studies show that organizational effects of androgens on cognition occur at blood levels far below those needed to influence reproductive morphology, which may be the reason that within-sex variation in sexually dimorphic cognitive skills is larger than between the sexes. After the first two months following birth, there is a significant overlap in the testosterone levels of boys and girls prior to puberty.

The perspective of assuming a computational bias in the study of cognitive sex differences may aid in understanding the scaffolding role of dorsal and ventral stream processing for the development of individual differences. Toward this end, we have developed simple, low-level computer based cognitive tasks that require vector estimation of a moving ball (dorsal stream) or recognition of a transient color change in a moving ball (ventral stream). A neurobehavioral picture of processing can be provided by participant’s reaction time and accuracy while using fMRI and/or mu suppression to measure neural activation during task performance.  Preliminary results show large sex differences in vector estimation accuracy and reaction time. Males are significantly more accurate than females, accompanied by a significantly longer reaction time. Activation of the MNS during the task is also significantly greater in males while they are doing the task. In contrast, females are significantly more accurate in detecting a color change in a moving ball than males. Studies have not been conducted on MNS activation during this task.  Overall, we view these studies as the foundation of an approach that may help us to gain further insight into functional neural substrates underlying individual cognitive differences, regardless of gender.

  

REFERENCES

1.  Maccoby EE, Jacklin CN. 1974. The Psychology of Sex Differences. Volumes 1 and 2. Stanford University Press, Stanford, CA.

2.  Arnold AP, Gorski RA. 1984. Gonadal steroid induction of structural sex differences in the central nervous system. Annu. Rev. Neurosci. 7:413-442.

3.  Ellis l, Hershberger s, Field E, Wersinger S, Pellis S, Geary D, Palmer C, Hoyenga K, Hetsroni A, Karadi K. 2008. Sex Differences: Summarizing More Than a Century of Research. Psychology Press, Taylor & Francis, Oxford.

4.  Ungerleider L,G., Mishkin M. 1982. Two cortical visual systems. In: Analysis of Visual Behavior (Ingle DJ et al., Eds), pp. 549-586, MIT Press, Cambridge, MA.

5.  Rao SC, Rainer G, Miller EK. 1997.  Integration of what and where in the primate prefrontal cortex.  Science 276:821-824.

6. Hickok G, Poeppel D. 2004. Dorsal and ventral streams: a framework for understanding aspects of the functional anatomy of language. Cognition 92:67-99.

7. Goodale MA, Milner, AD. 1992.  Separate visual pathways for perception and action. Trends in Neuroscience 15:20–25.

8. Butler T, Imperato-McGinley J, Pan H, Voyer D, Cordero J, Zhu YS, Stern E, Silbersweig D. 2006. Sex differences in mental rotation: top-down versus bottom-up processing. Neuroimage 32:445-456.

9. Chai XJ, Jacobs LF. 2009. Sex differences in directional cue use in a virtual landscape. Behav. Neurosci. 123:276-283.

10. Pineda, J. A. 2005. The functional significance of mu rhythms: translating "seeing" and "hearing" into "doing". Brain Res. Brain Res. Rev. 50:57-68.

11. Gallese V. 2007. The “conscious” dorsal stream: Embodied simulation and its role in space and action conscious awareness. Psyche 7:1-20.

12. Handa, RJ, McGivern, RF.   2002. Androgen Receptors in the Brain:  A behavioral perspective. In:  Androgens and Androgen Receptor: Mechanisms, Functions and Clinical Applications. CC Chang, Ed. Kluwer Academic/Plenum Publishers, NY, NY.