Vision begins with the image focused on the retina at the back of the eye. One way of examining the first stage of processing of that image is to measure ability to detect elementary changes in light and dark across the image; because the image probably also varies with time, we also measure the effects of different rates of change, to approximate changes that might occur in real-world viewing. We are dealing here with the effects of sex on these elementary visual sensations.
At least one sex-effect is well-known. Color vision depends on three types of cone photoreceptors in the retina: some are more sensitive to the longer wavelengths of light (L-cones), some to the middle wavelengths (M-cones), and some to the shorter wavelengths (S-cones). The genes coding for two of these cone photoreceptors (L- and M-cones) are carried on the X-chromosome and any malformation of either gene in a female is necessarily expressed in the phenotype of a male offspring who inherits that gene. However, it seems to be tacitly assumed that there are no other major sex effects on visual capacities. (But see companion paper [1
], for sex effects on color appearance.) Of course, major clinically oriented surveys that deal mostly with acuity and optical issues test large samples from a population and examine many variables, including sex. For example: this was done in the Hispanic Health and Nutrition Examination Survey of Hispanic and non-Hispanic populations – in fact this study found no significant sex-related acuity differences in children and adolescents [2
]. However, other studies of visual acuity have reported significant sex differences in adults [3
Despite mandates of major granting agencies to include, where possible, equal numbers of males and females in studies, little attention is paid to sex differences. For example, we examined all papers published in 2010 in a major journal, Vision Research. We tabulated all studies that used psychophysical or physiological measures to test humans’ basic capacities, such as spatial and temporal resolution, stereopsis, motion detection, and so on. We identified 410 such studies, of which only 96 (23.4%) broke down their samples into males and females. Also, at least in this sample of publications, only 4 studies (less than 1%) used sex as a comparison variable.
The lack of attention to sex differences in vision is surprising given the considerable body of work comparing male and female visually based perceptual and cognitive abilities: for example, females are said to make many more fine distinctions among colored objects [5
]. But if the initial information is processed differently according to sex, at least some of the “higher” differences might be due to differences in the results of this processing. Despite this possibility, little attention has been lavished on sex differences in basic visual sensations that are determined at “early” stages of the eye and brain.
Sex differences have been examined in other sensory systems: all showed clear sex differences. In the auditory system, there were differences in electrical responses of the brain (auditory evoked potentials); also ears of females were more likely to produce spontaneous emissions of sounds from the ear (otoacoustic emissions), and in those females, hearing sensitivity was better than in males. All these differences could be related to the masculinizing effects of androgens [7
]. Similarly, for the olfactory system, it seems clear that, in most cases, females had better sensitivity, and discriminated and categorized odors better than males [10
]. The general conclusion is that for audition and olfaction, as well as taste and somato-sensory sensitivity, females have greater sensitivity than males [12
The substrate for these sex differences may be linked directly to gonadal steroid hormones: in rhesus monkeys, large numbers of androgen receptors are found on neurons throughout the cerebral cortex, including visual cortex. This androgen binding “may have considerable impact on cortical functioning in primates at postnatal as well as prenatal ages” [13
]. There are similar findings for rats, in whom males have more androgen receptors than females, and these are especially plentiful in primary visual cortex [14
]. A recent review has reiterated these findings and concluded that in both humans and rats the largest concentration of androgen receptors in the forebrain is in the cerebral cortex and not the hypothalamic and limbic areas associated with reproduction [15
]: these findings might be general across mammals. All the above authors strongly emphasize that the distribution of androgen binding receptors may have considerable impact on cortical development and maturation of visual functioning.
Furthermore, in rats, androgens, but not estrogen, directly modify development of the visual cortex. Androgens reduce the early post-natal cell-death (apoptosis) of the visual cortex; as a consequence males have 20% more neurons in the visual cortex [16
]. This organizational effect is androgen-specific: early exposure of female rats to androgens (implanted capsules of dihydrotestosterone) led to these effects; early exposure to estrogen (implanted capsules of estradiol) did not inhibit post-natal cell-death [16
Factors other than gonadal receptors may also be involved [18
]. Theoretically, females might have a double “dose” of sex-related genes. To compensate for this, one of each pair of X-chromosomes is silenced [19
]. Many humans have multiple L and M genes – we are polymorphic for these genes [20
]. As a consequence, different retinal areas might express different alleles, affecting the responses of these areas and the brain sites associated with different retinal areas. Moreover, the X-chromosome may have a loading of “male-benefit” genes: thus, any recessive alleles must, of necessity, be expressed in a male [18
]. Furthermore, some of the sex effects we report here could be either organizational or activational and could depend on estrogen rather than testosterone; they could even be due to other sex-related genes [22
In short, it seems highly unlikely that vision will differ from other sensory modalities that show sex differences in function. It seems parsimonious to assume that the plethora of androgen receptors and androgen effects in primary visual cortex exert a measurable and important impact on visual sensory capacities.
We study differences between males and females in basic visual functions. The work we report here is not directly hypothesis-driven – it is exploratory. Our weak hypothesis is based on the fact that males have higher levels of androgens and that there are large numbers of testosterone receptors in visual cortex: therefore we anticipate sex differences. We cannot formulate a strict hypothesis because we cannot manipulate testosterone levels, nor have we measured them. This problem is endemic to most studies of sex differences in adult populations and is exacerbated by the possibility that some of the differences might not even be androgen driven.
We have developed a Battery of Visual Tests designed to study different visual sensory capacities, each with state-of-art precision. Tests include: (i) ability to resolve/detect patterns that vary in space (from coarse to fine detail ) and time (from very fast to very slow rates of change) -- these measures determine the spatio-temporal contrast-sensitivity function; (ii) detecting a small offset between two lines (vernier acuity); (iii) motion detection; (iv) binocular depth perception (stereopsis); and (v) color vision -- especially color appearance (see [1
]). One of the reasons for choosing these specific tests is that each emphasizes a different level and locus in the central nervous system. Specifically, the contrast-sensitivity function probably depends on the responses of neurons in primary visual cortex (discussed below); motion detection probably depends on responses of neurons in area MT of the temporal lobe; and some aspects of color appearance depend on an intact infero-temporal cortex (e.g., [23
In this paper we report on sex differences in spatio-temporal contrast-sensitivity. Spatial contrast-sensitivity (S-CSF) describes the visual system’s ability to perceive changes in brightness across space, as in reading a letter on an eye chart, or recognizing a face. The details of our procedures are given below.
The protocol for each measure in our Battery of Visual Tests is rigidly controlled. A consequence is that the thresholds we record from individuals now are precisely comparable to those recorded from individuals tested years ago. To minimize run-time errors, either the apparatus as a whole or the run-time stimulus sequence and data collection are computer-controlled. When these thresholds are measured, a participant adopts an internal criterion for reporting when a stimulus is visible; unfortunately these criteria vary across time within and among participants. Where possible, therefore, we use criterion-free psychophysical procedures: the stimulus is presented in one of two alternative positions or time intervals; after each presentation participants must choose where the stimulus occurred, regardless of confidence in their choice; stimulus strength is reduced from trial to trial until correct responding reaches chance levels. This “forced-choice” procedure is criterion free. In practice, we use a modified Bayesian approach to vary the stimuli in order to find threshold; trials are continued until the estimates of threshold reach an asymptote at a specified (99%) statistical confidence level (QUEST algorithm; [25
]). Also, we attempt to test each participant on the entire battery of tests to examine possible interactions among different capacities – those capacities whose thresholds are correlated presumably share a neuronal substrate.
We have amassed large databases from participants who participated in most or all of the Battery’s tests. We now have sufficient data to examine possible sex effects. Even if effect magnitudes are small or subtle, they cannot be ignored because they point to our hypothesized developmental and maturational effects of gonadal hormones on very basic sensory functions.
The following describes how, in the laboratory, we measure the S-CSF: we use grating patterns (alternating light and dark bars); the gratings vary from broad, coarsely spaced bars to fine, closely spaced bars. On each trial the grating pattern is either horizontal or vertical and the participant must choose between these alternative orientations. The spacing of the bars is measured as the number of light–dark cycles within one degree of visual space (cy/deg). The change from light to dark bars follows a sinusoidal profile in which intensity is varied symmetrically above and below a mean gray level; the difference between maximal and minimal intensities of the pattern is reduced until the pattern is no longer detectably different from a uniform gray – that is, the grating’s contrast is reduced until the pattern is at threshold.
The gratings had sinusoidal profiles because any image can be synthesized from (is equivalent to) a specific set of sinusoids of different frequencies, amplitudes, and phases (Fourier’s theorem). Limiting measures of spatial resolution only to the finest detail that can be perceived (acuity) is akin to testing auditory capacities using only the top end of the piano keyboard. High spatial frequencies are indeed important for discriminating among images that are similar to each other. However, sensitivity to the lower spatial frequencies greatly influences our ability to recognize and categorize parts of the image that refer to the diverse objects on the visual field [26
Similarly, temporal contrast-sensitivity (T-CSF) refers to the visual system’s ability to perceive changes in brightness over time, as when the stimulus flickers, or the retinal image slips across the retina due to object and/or viewer motion. However, S- and T-CSFs are strictly non-separable functions: the precise function that is obtained depends on the values of both sets of parameters – these define an entire three-dimensional sensitivity surface. Kelly, an early advocate of this non-separability [29
], used gratings of each spatial frequency which moved horizontally at a fixed velocity. This was impractical in our situation so we chose to measure the entire spatio-temporal (ST-CSF) surface using a series of spatial gratings, each of whose contrasts was modulated at each of a fixed series of temporal rates. The time-profile of the modulation was sinusoidal. The units for these temporal rates are variation-cycles/s (Hz).