Elisabeth Stöttinger

Division of labour within the visual system: fact or fiction? Which kind of evidence is appropriate to clarify this debate?

The perception versus action hypothesis of Goodale and Milner (Trends Neurosci 15:20–25, 1992) and Milner and Goodale (The visual brain in action. Oxford University Press, Oxford, 1995) postulated two different pathways within the visual system—one for action and one for perception. With the help of pictorial illusions, evidence for this dissociation was found in various studies. There is an ongoing debate, however, as to whether or not this evidence is biased by methodological issues. Indeed, relevant and decisive data can come only from those studies that (1) match conditions appropriately with respect to task demands, (2) use illusions that do not provide any potential obstacles for the hand, (3) do not risk that grasping is either memory driven (when the target is not visible) or online corrected (due to a direct comparison of the grip aperture with the size of the target object), (4) do not confound differences between perception and action conditions with differences in visual feedback, and (5) correct for differences in response functions between grasping and perception. In following all these points outlined above we found support for the perception versus action hypothesis: grip aperture follows actual size independent of illusory effects, while perceived length as indicated by finger–thumb span clearly was subject to the illusion.

Getting a grip on illusions: replicating Stöttinger et al [Exp Brain Res (2010) 202:79–88] results with 3-D objects

Studies using visual illusions to demonstrate a dissociation within the visual system can provide relevant and decisive data only if certain methodological points are taken into account. Although, our previous work (Stöttinger et al. in Exp Brain Res 202:88–97, 2010) followed these points, the task made use of only 2-D stimuli which may raise doubts concerning the nature of grasping in that experiment. We therefore replicated the study using a 3-D version of the empty space illusion. Consistent with the earlier study, that used 2-D stimuli, we found that grip aperture followed actual target size independent of illusory effects, while perceived length, as indicated by finger-thumb span, clearly was subject to the illusion. Therefore, the prior results cannot be due to the use of 2-D stimuli. Together, these two studies provide clear evidence for the perception versus action hypothesis.

Grasping the diagonal: Controlling attention to illusory stimuli for action and perception

Since the pioneering work of [Aglioti, S., DeSouza, J. F., & Goodale, M. A. (1995). Size-contrast illusions deceive the eye but not the hand. Current Biology, 5(6), 679-685] visual illusions have been used to provide evidence for the functional division of labour within the visual system-one system for conscious perception and the other system for unconscious guidance of action. However, these studies were criticised for attentional mismatch between action and perception conditions and for the fact that grip size is not determined by the size of an object but also by surrounding obstacles. Stoettinger and Perner [Stoettinger, E., & Perner, J., (2006). Dissociating size representations for action and for conscious judgment: Grasping visual illusions without apparent obstacles. Consciousness and Cognition, 15, 269-284] used the diagonal illusion controlling for the influence of surrounding features on grip size and bimanual grasping to rule out attentional mismatch. Unfortunately, the latter objective was not fully achieved. In the present study, attentional mismatch was avoided by using only the dominant hand for action and for indicating perceived size. Results support the division of labour: Grip aperture follows actual size independent of illusory effects, while finger-thumb span indications of perceived length are clearly influenced by the illusion.

Statistical and perceptual updating: correlated impairments in right brain injury

Abstract It has been hypothesized that many of the cognitive impairments commonly seen after right brain damage (RBD) can be characterized as a failure to build or update mental models. We (Danckert et al. in Neglect as a disorder of representational updating. NOVA Open Access, New York, 2012a; Cereb Cortex 22:2745–2760, 2012b) were the first to directly assess the association between RBD and updating and found that RBD patients were unable to exploit a strongly biased play strategy in their opponent in the children’s game rock, paper, scissors. Given that this game required many other cognitive capacities (i.e., working memory, sustained attention, reward processing), RBD patients could have failed this task for various reasons other than a failure to update. To assess the generality of updating deficits after RBD, we had RBD, left brain-damaged (LBD) patients and healthy controls (HCs) describe line drawings that evolved gradually from one figure (e.g., rabbit) to another (e.g., duck) in addition to the RPS updating task. RBD patients took significantly longer to alter their perceptual report from the initial object to the final object than did LBD patients and HCs. Although both patient groups performed poorly on the RPS task, only the RBD patients showed a significant correlation between the two, very different, updating tasks. We suggest these data indicate a general deficiency in the ability to update mental representations following RBD.

Spatial biases in number line bisection tasks are due to a cognitive illusion of length

Placing arrow heads (Judd Illusion) or numbers of different magnitude at the end of a line biases perception of the centre of the line. For the Judd Illusion, it is known that this bias depends on the method used: a deliberate (more perceptually based) marking of the centre with a pen is more subject to the illusion than are fast (more action-based) ballistic pointing movements made towards the centre. It has been suggested that the number bias also reflects a cognitive illusion of length. To test this assumption, we used two different response methods in line bisection tasks while lines were flanked by arrow heads or numbers of different magnitudes. For both conditions, we found that the more action-based response method showed less bias. Since the pattern of biases induced by flanking numbers and arrow heads are similar, we confirm that the spatial bias produced by numerical magnitude reflects a cognitive illusion of length.

Assessing perceptual change with an ambiguous figures task: Normative data for 40 standard picture sets

In many research domains, researchers have employed gradually morphing pictures to study perception under ambiguity. Despite their inherent utility, only a limited number of stimulus sets are available, and those sets vary substantially in quality and perceptual complexity. Here we present normative data for 40 morphing picture series. In all sets, line drawings of pictures of common objects are morphed over 15 iterations into a completely different object. Objects are either morphed from an animate to an inanimate object (or vice versa) or morphed within the animate and inanimate object categories. These pictures, together with the normative naming data presented here, will be of value for research on a diverse range of questions, from perceptual processing to decision making.

Sequential Decisions: A Computational Comparison of Observational and Reinforcement Accounts

Right brain damaged patients show impairments in sequential decision making tasks for which healthy people do not show any difficulty. We hypothesized that this difficulty could be due to the failure of right brain damage patients to develop well-matched models of the world. Our motivation is the idea that to navigate uncertainty, humans use models of the world to direct the decisions they make when interacting with their environment. The better the model is, the better their decisions are. To explore the model building and updating process in humans and the basis for impairment after brain injury, we used a computational model of non-stationary sequence learning. RELPH (Reinforcement and Entropy Learned Pruned Hypothesis space) was able to qualitatively and quantitatively reproduce the results of left and right brain damaged patient groups and healthy controls playing a sequential version of Rock, Paper, Scissors. Our results suggests that, in general, humans employ a sub-optimal reinforcement based learning method rather than an objectively better statistical learning approach, and that differences between right brain damaged and healthy control groups can be explained by different exploration policies, rather than qualitatively different learning mechanisms.

Consistency in exchange for inappropriately matched visual feedback? A comment on Franz and Gegenfurtner (2008) "Grasping visual illusions: consistent data and no dissociation".

Franz and Gegenfurtner (2008) argue that the evidence for a division of labour within the visual system for action and perception is flawed because perception is often measured by manual estimation, which responds in general with a larger slope to a change of physical size than does adjusting. Therefore results obtained under manual estimation have to be corrected for this difference in slope: In a reanalysis of six studies grasping and perception were equally influenced by the illusion after this correction. However, closer inspection of methods reveals that visual feedback was confounded with conditions (suppressed vision while grasping vs. full vision while adjusting). We argue that studies can produce relevant and decisive data only when they (a) do not confound conditions with visual feedback, (b) do not allow online corrections of the action due to a direct comparison of the hand with the target, and (c) do not provide any risk of grasping being memory driven when the target is removed.

The Neural Systems for Perceptual Updating

ABSTRACT In a constantly changing environment we must adapt to both abrupt and gradual changes to incoming information. Previously, we demonstrated that a distributed network (including the anterior insula and anterior cingulate cortex) was active when participants updated their initial representations (e.g., its a cat) in a gradually morphing picture task (e.g., now its a rabbit; Stöttinger et al., 2015). To shed light on whether these activations reflect the proactive decisions to update or perceptual uncertainty, we introduced two additional conditions. By presenting picture morphs twice we controlled for uncertainty in perceptual decision making. Inducing an abrupt shift in a third condition allowed us to differentiate between a proactive decision in uncertainty‐driven updating and a reactive decision in surprise‐based updating. We replicated our earlier result, showing the robustness of the effect. In addition, we found activation in the anterior insula (bilaterally) and the mid frontal area/ACC in all three conditions, indicative of the importance of these areas in updating of all kinds. When participants were naïve as to the identity of the second object, we found higher activations in the mid‐cingulate cortex and cuneus – areas typically associated with task difficulty, in addition to higher activations in the right TPJ most likely reflecting the shift to a new perspective. Activations associated with the proactive decision to update to a new interpretation were found in a network including the dorsal ACC known to be involved in exploration and the endogenous decision to switch to a new interpretation. These findings suggest a general network commonly engaged in all types of perceptual decision making supported by additional networks associated with perceptual uncertainty or updating provoked by either proactive or reactive decision making. HighlightsThe world is in flux – we constantly need to update our interpretations.Things can change abruptly or gradually – requiring different processes.Frontal brain regions are engaged in both types of perceptual decision making.While temporal‐parietal areas are specifically involved in proactive decision making.

Updating impairments and the failure to explore new hypotheses following right brain damage

We have shown recently that damage to the right hemisphere impairs the ability to update mental models when evidence suggests an old model is no longer appropriate. We argue that this deficit is generic in the sense that it crosses multiple cognitive and perceptual domains. Here, we examined the nature of this updating impairment to determine more precisely the underlying mechanisms. We had right (RBD, N = 12) and left brain damaged (LBD, N = 10) patients perform versions of our picture-morphing task in which pictures gradually morph from one object (e.g., shark) to another (e.g., plane). Performance was contrasted against two groups of healthy older controls, one matched on age (HCO-age-matched, N = 9) and another matched on general level of cognitive ability (HCO-cognitively-matched, N = 9). We replicated our earlier findings showing that RBD patients took longer than LBD patients and HCOs to report seeing the second object in a sequence of morphing images. The groups did not differ when exposed to a morphing sequence a second time, or when responding to ambiguous images outside the morphing context. This indicates that RBD patients have little difficulty alternating between known representations or labeling ambiguous images. Instead, the difficulty lies in generating alternate hypotheses for ambiguous information. Lesion overlay analyses, although speculative given the sample size, are consistent with our fMRI work in healthy individuals in implicating the anterior insular cortex as critical for updating mental models.