Does the Human Dorsal Stream Really Process Elongated Vegetables?

What do zucchini and hammers have in common? Both might be processed by the dorsal stream.

The primate visual system is divided into ventral ("what") and dorsal ("where") visual streams that are specialized for object recognition and spatial localization, respectively (Mishkin et al., 1983; Haxby et al., 1991).


Goodale and Milner (1992) conceptualized the two pathways as "vision for perception" and "vision for action":

We propose that the ventral stream of projections from the striate cortex to the inferotemporal cortex plays the major role in the perceptual identification of objects, while the dorsal stream projecting from the striate cortex to the posterior parietal region mediates the required sensorimotor transformations for visually guided actions directed at such objects.

Other researchers have extended the degree of specialization shown by the visual and semantic systems. Some studies have suggested there might be category-specific processing of living and non-living things (e.g., animals and tools), although the reasons for this specialization are a matter of debate (Caramazza & Shelton, 1998; Thompson-Schill et al., 1999). Chao and Martin (2000) found that pictures of tools activated the left posterior parietal cortex in the dorsal stream to a greater extent than pictures of animals, houses, and faces. The idea is that objects with salient motor-based properties (hammers) should recruit "vision for action" cortical regions to a greater extent than objects without such affordances (zucchini).

More recently, Almeida and colleagues (2008) used two different visual masking techniques in a priming study designed to isolate the influence of the dorsal stream:

We used two techniques to render prime pictures invisible: continuous flash suppression (CFS), which obliterates input into ventral temporal regions, but leaves dorsal stream processes largely unaffected, and backward masking (BM), which allows suppressed information to reach both ventral and dorsal stream structures.

Their results suggested that categorically related primes suppressed under CFS still facilitated reaction times to tool targets, but not to animal targets. In other words, participants were faster to classify tools when preceded by a picture of a tool than when preceded by a picture of an animal, and this priming effect held up when the ventral stream was unavailable.

A new study by Sakuraba et al. (2012) wanted to clarify which specific attributes of tools are processed by the dorsal stream, so they used a greater variety of categorically related and unrelated prime stimuli suppressed under CFS, as shown below.

Fig. 1 (Sakuraba et al., 2012). Procedure using CFS. Different images were presented into the subject's left and right eyes by using anaglyphs. Dynamic high-contrast random-noise patterns (10 Hz) were presented to the dominant eye, while low-luminance, low-contrast prime stimuli were presented to the nondominant eye. Subjects could report the dynamic noise but not the static image. Each trial started with a fixation cross for 500 ms, followed for 200 ms by a prime stimulus suppressed by CFS. Finally, a target stimulus masked by 70% additive noise was presented until the subject responded (maximum duration: 3 s) by pressing a key to indicate the category of the target stimulus.

One of the manipulated attributes was shape. Non-elongated ("stubby") tools, elongated vegetables, and stubby vegetables¹ were used as primes for elongated tools (e.g., hammer, ax, fork, etc.). Other conditions used geometric shapes as primes.

In Experiment 2, we used tool pictures without elongated shape components, namely stubby tools (e.g., a punch, a squeezer, a mouse, and so on). ... In Experiment 3, elongated stick figures were used as prime stimuli. In Experiment 4, elongated and stubby vegetable pictures were presented as prime stimuli. Because the elongated shapes involve an orientation component, we could not exclude the possibility that orientation, rather than shape attribute, explained the results. Therefore, we conducted Experiment 5 to clarify this. We used elongated stick figures, diamond shapes, and cut circles that were rotated in 45° increments as prime stimuli.

Interestingly, membership in the category of tools per se was irrelevant; it was the shape of the prime that mattered.

Fig. 4 (Sakuraba et al., 2012). Priming effect in Experiment 4 and 5.² Light and dark gray bars represent mean priming effects to tool targets and animal targets, respectively. Error bars indicate SEM. The pictures represent examples of the prime stimuli we used.

This throws a wrench (so to speak) into the dorsal stream as the vision for action pathway, unless you normally use a zucchini to pound your nails into the wall. The less dramatic interpretation is that the categorical information obtained by viewing pictures of tools isn't neatly respected by the dorsal stream, but visually-guided reaching and grasping remain unscathed.

Footnote

¹ The "stubby vegetables" were my favorite part of the paper.

² You might want to quibble with the size and functional significance of the priming effect. Although statistically significant, it was rather small.

References

Almeida J, Mahon BZ, Nakayama K, Caramazza A. (2008). Unconscious processing dissociates along categorical lines. Proc Natl Acad Sci. 105:15214-18.

Caramazza A, Shelton JR. (1998). Domain-specific knowledge systems in the brain: the animate-inanimate distinction. J Cogn Neurosci. 10:1-34.

Chao LL, Martin A. (2000). Representation of manipulable man-made objects in the dorsal stream. Neuroimage 12:478-84.

Goodale MA, Milner AD. (1992). Separate visual pathways for perception and action. Trends Neurosci. 15:20-5.

Haxby JV, Grady CL, Horwitz B, Ungerleider LG, Mishkin M, Carson RE, Herscovitch P, Schapiro MB, Rapoport SI. (1991). Dissociation of object and spatial visual processing pathways in human extrastriate cortex. Proc Natl Acad Sci 88:1621-5.

Mishkin M, Ungerleider LG, Macko KA (1983). Object vision and spatial vision: two cortical pathways. Trends Neurosci. 6:414-7.

Sakuraba S, Sakai S, Yamanaka M, Yokosawa K, & Hirayama K (2012). Does the human dorsal stream really process a category for tools? The Journal of Neuroscience, 32 (11), 3949-53. PMID: 22423115

Thompson-Schill SL, Aguirre GK, D'Esposito M, Farah MJ. (1999). A neural basis for category and modality specificity of semantic knowledge. Neuropsychologia 37:671-6.

The Art of Manliness on How to Handle a Hammer.