Methodologically, investigating spatial discrimination involves roughly the same sort of process that is involved in investigating the discrimination of qualitative features such as colour and pitch. That is, one begins by doing psychophysics in order to test the limits of spatial discrimination and to determine the structure or topography of phenomenological space. Then, we can turn to neuroscience to look for the mechanisms which can explain the psychophysical data.
For example, in testing the spatial resolution of the somatosensory system, a common type of task that has traditionally been used is the so-called “two-point discrimination” task, in which an experimenter touches a subject on the skin with either one or two sharp points. The subject must report whether one point or two points were used, and the smallest distance at which the subject can discriminate one point from two at a level sufficiently above chance is known as the “two-point threshold”. (Of course, this
threshold varies across different bodily locations; at the fingertips and lips, the threshold is as small as 2mm, whereas on the back the threshold can be as high as 40mm.) The addition of other psychophysical measures can provide an even more accurate picture of tactile spatial resolution; for example, the two-point threshold can be supplemented with an orientation discrimination task which measures the ability of the subject to
discriminate the alignment of the points (Tong, Mao, & Goldreich, 2013). Similarly, experimenters can also test the ability of subjects to detect the orientation of a grating that is touching the skin (Johnson & Phillips, 1981). The resulting (psychophysically-
determined) picture of somatosensory space that develops is the well-known
‘homunculus’: the grotesquely disproportionate representation of the human body that has excessively large hands, lips, and face relative to the rest of the body.89 Importantly,
89 It might seem as though this resulting spatial representation of the body’s surface is somehow inaccurate;
a distortion of its spatial structure. However, a better way to think of it is as defining the ‘density’ of somatosensory space, or the ‘grain of resolution’ of our representation of different body parts. Moreover, sensory systems are rarely designed to represent absolute values of stimuli. Rather, as Akins (1996) puts it, sensory systems are designed to be “narcissistic”; to solve particular informational problems, not to represent veridically; i.e., systematically encoding the absolute value of external relations without embellishment.
the proportions of the somatosensory homoculus can be directly explained by
neuroscience; specifically, they are a direct result of both the number of sensory receptors in any particular location on the body’s surface, and the amount of neural real estate in the somatosensory cortex (and motor cortex) dedicated to processing the signals from those receptors.
Similarly, the unique structural topography of auditory space that is discovered by psychophysics can be explained by the neural mechanisms that underlie our ability to localize sound stimuli in our environment. Contemporary methods of mapping auditory space rely on varying a number of independent factors (such as the azimuth, elevation, and distance of the sound source; the frequency, pitch, intensity, and timber of the sound; the nature of the environment; and so on) along some particular dimension(s) and
observing the subsequent effects on the subject’s ability to discriminate the spatial location of a sound source. The resulting picture of auditory space has a very specific topographic structure: For example, spatial localization in audition is vulnerable to a particular kind of illusion known as front-back reversals, wherein sound stimuli that are located directly in front of the head and directly behind the head are apt to be confused with one another. 90 The explanation for this phenomenon has to do with the fact that
auditory information about the azimuth of a sound source is based mainly on calculating certain types of subtle differences between a sound as it arrives at each ear. When the stimulus is directly in front of or directly behind the listener, those types of differences (which I will discuss in greater detail below) are equivalent, thus leading to front-back reversal. Similarly, humans are much, much worse at estimating the egocentric distance of a sound source than its azimuth or elevation. Again, this can also be explained by the particular processing structure of neural mechanisms that underlie auditory localization (which again, I will discuss in greater detail below). Finally, the accuracy of human auditory localization depends in many ways on the particular spectral frequencies of the
90 This effect—and indeed, psychophysical investigation of auditory spatial localization in general—has
been studied since as early as 1796, when Giovanni Venturi demonstrated that subjects could point in the direction of a sound source. (Venturi demonstrated this by having subjects sit in an open field blindfolded, while he walked around them, intermittently playing notes on his flute and asking subjects to point in the direction of the sound.)
sound, which again is a result of the sensitivity and excitability of particular cell types in the auditory system to sounds in a certain narrow frequency range.
In any case, the general point should now be clear: methodologically speaking, providing an account of the spatial phenomenology of sensory experience requires us to first do psychophysical experiments to determine the topographic structure of the represented spatial relations in a given modality, and then do neuroscience to discover neural mechanisms which can explain that structure. Of course, thus far I have only discussed the uni-modal representation of space; however, one can also do cross-modal experiments in both psychophsyics and neuroscience in order to discover the effects that spatial
representations in one modality have on representations of spatial relations in a different modality. And indeed, as I shall argue below, this sort of cross-modal investigation is absolutely crucial to the project of providing an account of the phenomenology of spatial experience, because the sensory representation of spatial location is inherently multi- modal in nature. However, for now this preliminary characterization of the methodology will suffice. In what follows, I will focus primarily on the neural underpinnings of spatial localization in order to provide an (admittedly oversimplified) explanation for the
structure of spatial phenomenology.