Learning to detect but not to grasp suppressed visual stimuli

Highlights

• Implication of the two-visual-systems hypothesis was tested in healthy participants.

• Can the dorsal system use invisible information, whereas the ventral system cannot?

• Participants grasped for invisible objects and reported their properties verbally.

• Chance performance but decreasing perceptual thresholds in grasping task.

• Perceptual performance possibly enhanced by concurrent visuomotor behavior.

Abstract

A central implication of the two-visual-systems hypothesis (TVSH) is that the dorsal visuomotor system (vision-for-action) can make use of invisible information, whereas the ventral system (vision-for-perception) cannot (Milner & Goodale, 1995). Therefore, actions such as grasping movements should be influenced by invisible information while conscious reports remain unaffected. To test this assumption, we used a dichoptic stimulation technique – continuous flash suppression (CFS) – which has the potency to render stimuli invisible for up to seconds (Tsuchiya & Koch, 2005). In two experiments using CFS, participants were asked to grasp for invisible bars of different sizes (Experiment 1) or orientations (Experiment 2), or to report both measures verbally. Target visibility was measured trial-by-trial using the perceptual awareness scale (PAS). We found no evidence for the use of invisible information by the visuomotor system despite extensive training (600 trials) and the availability of haptic feedback. Participants neither learned to scale their maximum grip aperture to the size of the invisible stimulus, nor to align their hand to its orientation. Careful control of stimulus visibility across training sessions, however, revealed a robust tendency towards decreasing perceptual thresholds under CFS. We discuss our results within the framework of the TVSH and with respect to alternative models which emphasize the close functional interaction between the dorsal and ventral visual systems.

Introduction

The ‘two visual systems’ hypothesis (TVSH) proposed by Milner and Goodale (1995) remains one of the most influential models of visual processing in the primate brain. It assumes a dissociation between consciously accessible ‘vision-for-perception’, mediated by the ventral cortical pathway, and consciously inaccessible ‘vision-for-action’, mediated by the dorsal pathway (Milner, 2012). One of the primary pillars this model rests on are the studies performed on patient D.F., who suffers from visual form agnosia as a result of extensive bilateral ventral stream damage following carbon monoxide poisoning (James et al., 2003, Milner et al., 1991; also see: Karnath, Rueter, Mandler, & Himmelbach, 2009). D.F. can successfully grasp for objects while she is unable to perform perceptual judgements such as size or shape estimation on them and her remaining visual capacities have been attributed to dorsal stream processes (Culham et al., 2003, Milner and Goodale, 2008).

In neurologically healthy participants, the demonstration of perception-action dissociations has proven to be difficult. A first study that showed that grasping (taken as a measure for dorsal processing) may be immune to the Ebbinghaus size illusion while perceptual measures were not (Aglioti, DeSouza, & Goodale, 1995) has prompted a multitude of studies, many of which do not support a dissociation (e.g., Franz, Gegenfurtner, Bülthoff, & Fahle, 2000; for reviews see Franz and Gegenfurtner, 2008, Westwood and Goodale, 2011). Another way to probe this dissociation in healthy participants has been paved by the introduction of a new technique called continuous flash suppression (CFS) which can render stimuli invisible for up to several seconds by flashing high-contrast images to one eye while showing the target stimulus to the other eye (Tsuchiya and Koch, 2005, Tsuchiya et al., 2006).

Evidence from behavioral priming experiments (Almeida et al., 2008, Sakuraba et al., 2012) and neuroimaging studies (Sterzer et al., 2008, Troiani et al., 2012) shows that stimulus information suppressed from awareness with CFS can reach higher-order visual areas and influence behavior.

It has been suggested that CFS could be used to “isolate” dorsal visual processing, i.e., to leave nonconscious visuomotor processes mediated by the dorsal stream intact while disrupting conscious perception mediated by the ventral stream (Fang and He, 2005, Lin and He, 2009; but see: Hesselmann and Malach, 2011, Hesselmann, 2013). If this was the case, one could expect participants to show above chance performance in grasping invisible stimuli, while not being able to report the stimuli's features verbally. Such behavior would classify as a form of “blindsight”, a rare dissociation in patients with lesions to primary visual cortex (V1), typically defined as residual forced-choice visual function in the absence of subjective awareness (Cowey, 2004, Stoerig and Cowey, 2007). According to the taxonomy for different blindsight subtypes proposed by Danckert & Rossetti, accurately acting upon blind field stimuli based on preserved activity in the dorsal stream classifies as “action-blindsight” (Danckert & Rossetti, 2005). We reasoned that testing for a similar perception-action dissociation in healthy observers using CFS would constitute a critical test for the TVSH, in particular its claim that ventral stream processes are associated with visual awareness, while dorsal processes are not. In this study, we aimed to generate a blindsight-like situation in normal observers to test whether information unavailable to conscious report would still be processed in a way to enable grasping for objects of different sizes (Experiment 1) or orientations (Experiment 2).