Research ReportIt's not all about looks: The role of object shape in parietal representations of manual tools
Introduction
The outstanding manual dexterity and capability of fine-motor coordination required for the use of manual tools renders humans unique among primates. Similarly to other evolutionary relevant categories—such as faces and bodies (for social interactions) and places (for spatial navigation)—the category of tools has been shown to activate a distinct network of brain areas. This tool-selective network comprises areas along the ventral (Bracci, Cavina-Pratesi, Ietswaart, Caramazza, & Peelen, 2012; Chao & Martin, 2000; Downing, Chan, Peelen, Dodds, & Kanwisher, 2006; Macdonald & Culham, 2015; Perini, Caramazza, & Peelen, 2014) and dorsal (Boronat et al., 2005; Chao, Haxby, & Martin, 1999; Hermsdörfer, Terlinden, Mühlau, Goldenberg, & Wohlschläger, 2007; Kastner, Chen, Jeong, & Mruczek, 2017; Mruczek, von Loga, & Kastner, 2013; Peeters et al., 2009; Valyear, Cavina-Pratesi, Stiglick, & Culham, 2007) visual streams.
The critical role of these brain regions in processing tool knowledge and use is supported by evidence from lesion studies. These studies report selective impairments in patients' ability to manipulate (Koski, Iacoboni, & Mazziotta, 2002; Leiguarda & Marsden, 2000; Randerath, Goldenberg, Spijkers, Li, & Hermsdörfer, 2010) or verbally identify tools (Damasio, Grabowski, Tranel, Hichwa, & Damasio, 1996; Tranel, Damasio, & Damasio, 1997), in cases where the ability to manipulate or name other objects (e.g., chairs or animals) is left intact (for a review, see Johnson-Frey, 2004). A converging body of neuroimaging literature further corroborates these neurophysiological findings. While tool-selective areas are reported in the ventral object-recognition pathway (Beauchamp & Martin, 2007; Bracci, Cavina-Pratesi, Connolly, & Ietswaart, 2016; Chao et al., 1999), organization of the areas in the dorsal stream reflects functional support for tool-specific actions (Bracci & Op de Beeck, 2016; Brandi, Wohlschläger, Sorg, & Hermsdörfer, 2014; Gallivan, Adam McLean, Valyear, & Culham, 2013; Lewis, 2006). Importantly, even though parietal areas also encode information about graspability (Almeida, Mahon, & Caramazza, 2010; Konen & Kastner, 2008), and dorsal tool-selectivity has been suspected to solely reflect the areas' preferences for graspable features (Creem-Regehr & Lee, 2005), this has been refuted with various neuroimaging designs (e.g., Brandi et al., 2014; Mruczek et al., 2013; Valyear et al., 2007) and clinical studies (e.g., Randerath et al., 2010). Overall, it has been repeatedly shown that parietal areas encode grasp-related properties common to all manipulable objects separately from the idiosyncratic tool-specific action-related features (Brandi et al., 2014; Chen, Snow, Culham, & Goodale, 2017; Lewis, 2006; Valyear et al., 2007).
However, a debate has recently arisen as to whether this selectivity for tools is attributable to the proposed action properties, or if it rather reflects a plain selectivity for visual features (Sakuraba, Sakai, Yamanaka, Yokosawa, & Hirayama, 2012). Analogously to the recently renewed debate in object vision (Rice, Watson, Hartley, & Andrews, 2014; Watson, Hartley, & Andrews, 2014; Bracci et al., 2016; Proklova, Kaiser, & Peelen, 2016), it is possible that the observed neural effects for tools do not reflect the abstract property of category membership, but rather the visual properties that correlate highly with stimulus category—in this case, the distinct, elongated shape shared by most tool-objects (Almeida et al., 2014). This shape-based hypothesis was initially tested by Sakuraba et al. (2012), who found that elongated objects (e.g., an elongated vegetable) activated the alleged tool-related dorsal substrates. However, recent work by Chen et al. (2017) examined the shape-based account in more detail, and concluded that parietal areas process both object shape and action-related properties.
More generally, the long-prevailing notion of the dorsal pathway strictly supporting “vision-for-action” computations has recently been challenged from different perspectives. Accumulating evidence shows that the dorsal stream extracts features such as colour, size or shape (e.g., Freud, Culham, Plaut, & Behrmann, 2017; Konen et al., 2008) to support “vision-for-perception” that are independent of those generated in the ventral stream (Freud, Ganel, Shelef, Hammer, Avidan, & Behrmann, 2017). Parietal areas have also been shown to encode visual information adaptively, in accordance with the task requirements and behavioural goals (Xu, 2018; Bracci, Daniels, & Op de Beeck, 2017). Taken together, this undermines the established notion that object representations in parietal cortex are primary in the service of action; a more the general purpose of these representations might be to support behavioural goals, including both action and perception.
To test if object representational content in parietal cortex reflects action-specific properties per se, here we created and behaviourally validated a stimulus set that eliminates the effect of visual features, including object shape (Fig. 1), and delineates between objects that are a) manipulable and nonmanipulable (graspability control); and b) having different degrees of body-extension property (tool-hand effect). More specifically, manual tools are defined as extensors of the internal body schema when used to manually operate on other objects (Kastner et al., 2017); for example, when using a hammer to pound a nail, the hammer is functionally extending the boundaries of the hand by inducing changes to the body-schema (Maravita & Iriki, 2004). The degree of perceived body extension—how much an object is physically and functionally extending the body's boundaries (Bracci & Peelen, 2013)—is used here as a data-driven definition of tool-objects. By orthogonally manipulating object shape across object categories with different action components, this study allowed us to fully disentangle the action- and shape-related object representations, and examine the independent contribution of action-related properties to dorsal stream representational content.
Section snippets
Participants
For the fMRI study, 17 right-handed healthy volunteers with normal or corrected-to-normal vision (Mage = 28, SDage = 7.1, 10 females) were recruited. No participants were excluded, but due to technical issues a total of 5 runs were excluded (in 2 participants). For the behavioural validation, 9 healthy volunteers (mean age 25 [SD 3.1], 5 females) were recruited. Decisions about the total number of participants were taken prior to data collection and were based on results from previous
Results
The aim of the present study was to test whether the previously reported tool selectivity in human parietal cortex (Chao & Martin, 2000; Kastner et al., 2017; Mruczek et al., 2013; Valyear et al., 2007) reflects the representation of tool-specific, action-related information, or can instead be fully explained by shape properties idiosyncratic to tools (e.g., Sakuraba et al., 2012). To this end, we constructed and behaviourally validated a stimulus set composed of object categories that differ
Discussion
Watching images of tools is known to elicit a distinct pattern of neural activity in parietal visual areas, assumingly because tools entail a potential for action—an idiosyncratic feature related to their functional use and not shared with other manipulable objects (Chao & Martin, 2000; Lewis, 2006; Valyear et al., 2007). However, recent reports (Almeida et al., 2014; Sakuraba et al., 2012) suggested that tool-related activity in the dorsal visual stream might be explained solely by object
Future directions and conclusion
The aim of the present study was to elucidate the role of visual- and action-related features on neural activations in the two most commonly reported tool-related parietal regions, IPS and SPL. We showed that, when controlling for visual features and manipulability, these two regions both elicit a distinct representational pattern for tool-specific, action-related features. This does not preclude the possibility that regions along the dorsal visual stream also encode visual properties (Freud,
Author contributions
Karla Matić: Conceptualization, Investigation, Data curation, Formal analysis, Visualization, Writing - original draft, Writing - review & editing.
Hans Op de Beeck: Conceptualization, Supervision, Funding acquisition, Writing - review & editing.
Stefania Bracci: Conceptualization, Methodology, Formal analysis, Supervision, Funding acquisition, Writing - review & editing.
Open practices
The study in this article earned Open Materials and Open Data badges for transparent practices. Materials and data for the study are available at https://osf.io/yg9b4/.
Acknowledgements
This work was supported by the FWO (Postdoctoral Fellowship 12S1317N and Research Grant 1505518N) to S.B., and the European Research Council (ERC Grant 2011-StG-284101), a federal research action (IUAP-P7/11), the KU Leuven Research Council (C14/16/031), Hercules Foundation grant ZW11_10, and an Excellence of Science Grant (G0E8718N) to H.O.B.
References (66)
- et al.
Grounding object concepts in perception and action: Evidence from fMRI studies of tools
Cortex
(2007) - et al.
Distinctions between manipulation and function knowledge of objects: Evidence from functional magnetic resonance imaging
Cognitive Brain Research
(2005) - et al.
View-invariant representation of hand postures in the human lateral occipitotemporal cortex
Neuroimage
(2018) - et al.
Representational content of occipitotemporal and parietal tool areas
Neuropsychologia
(2016) - et al.
Representation of manipulable man-made objects in the dorsal stream
Neuroimage
(2000) - et al.
What role does “elongation” play in “tool-specific” activation and connectivity in the dorsal and ventral visual streams?
Cerebral Cortex
(2017) - et al.
Neural representations of graspable objects: Are tools special?
Cognitive Brain Research
(2005) - et al.
Human parietal cortex in action
Current Opinion in Neurobiology
(2006) - et al.
Assignment of functional activations to probabilistic cytoarchitectonic areas revisited
Neuroimage
(2007) The neural bases of complex tool use in humans
Trends in Cognitive Sciences
(2004)
A brief comparative review of primate posterior parietal cortex: A novel hypothesis on the human toolmaker
Neuropsychologia
Matching categorical object representations in inferior temporal cortex of man and monkey
Neuron
Do human brain areas involved in visuomotor actions show a preference for real tools over visually similar non-tools?
Neuropsychologia
Tools for the body (schema)
Trends in Cognitive Sciences
Against hyperacuity in brain reading: Spatial smoothing does not hurt multivariate fMRI analyses?
NeuroImage
Different left brain regions are essential for grasping a tool compared with its subsequent use
Neuroimage
Neural basis of pantomiming the use of visually presented objects
Neuroimage
A neural basis for the retrieval of conceptual knowledge
Neuropsychologia
Does tool-related fMRI activity within the intraparietal sulcus reflect the plan to grasp?
Neuroimage
Patterns of response to visual scenes are linked to the low-level properties of the image
Neuroimage
The role of the dorsal visual processing stream in tool identification
Psychological Science
Grasping with the eyes: The role of elongation in visual recognition of manipulable objects
Cognitive, Affective & Behavioral Neuroscience
Closely overlapping responses to tools and hands in left lateral occipitotemporal cortex
Journal of Neurophysiology
Task context overrules object- and category-related representational content in the human parietal cortex
Cerebral Cortex
Dissociations and associations between shape and category representations in the two visual pathways
Journal of Neuroscience
Body and object effectors: The organization of object representations in high-level visual cortex reflects body-object interactions
Journal of Neuroscience
The psychophysics toolbox
Spatial Vision
The neural correlates of planning and executing actual tool use
Journal of Neuroscience
FMRI reveals a dissociation between grasping and perceiving the size of real 3D objects
Plos One
Functional magnetic resonance imaging reveals the neural substrates of arm transport and grip formation in reach-to-grasp actions in humans
Journal of Neuroscience
Attribute-based neural substrates in temporal cortex for perceiving and knowing about objects
Nature Neuroscience
Functional magnetic resonance imaging during pantomiming tool-use gestures
Experimental Brain Research
Cytoarchitectonic identification and probabilistic mapping of two distinct areas within the anterior ventral bank of the human intraparietal sulcus
Journal of Computational Neurology
Cited by (2)
Action goals and the praxis network: an fMRI study
2022, Brain Structure and FunctionGraspability modulates the stronger neural signature of motor preparation for real objects vs. Pictures
2021, Journal of Cognitive Neuroscience