Shaping multisensory action–space with tools: evidence from patients with cross-modal extinction

Abstract

Recent findings from neurophysiology, neuropsychology and psychology have shown that peri-personal space is represented through an integrated multisensory processing. In humans, the interaction between peri-personal space representation and action execution can be revealed through the use of tools that, by extending the reachable space, modify the strength of visual–tactile extinction. We have previously shown that the peri-hand space whereby vision and touch are integrated can be expanded, and contracted, depending upon tool-use. Here, we show that these dynamic changes critically depend upon active tool-use, as they are not found after an equally long, but passive exposure to an elongated (hand + tool) body configuration. We also show that the extent of the peri-hand space elongation, as assessed at fixed far location (60 cm from the hand), varies according to the tool length such that a 30 cm long tool produced less elongation than a 60 cm long tool. This reveals for the first time that the distal border of elongated area is not sharply limited to the tool length, but extends beyond its physical size to include a peri-tool space whereby the strength of visual–tactile integration seems to fade. Remarkably, a similar amount of peri-hand space elongation was found when the effects of using a 30 cm long tool were compared with those produced by using a tool that was physically 60 cm long, but operationally 30 cm long. By dissociating with this ‘hybrid’ tool, the amount of space that is globally added to the hand (60 cm) from the one that is actually reachable (30 cm), we provide here the first evidence that the extent of peri-hand space elongation after tool use is tightly related to the functionally effective length of the tool, and not merely to its absolute length.

Introduction

Tools enable us to modify our action–space for various purposes, facilitating our daily interactions with objects in the environment (Beck, 1980, Napier, 1956). Also, non-human primates can spontaneously use tools for diverse purposes (e.g., branch-hook-use during locomotion and leaf-pads-use during feeding) and acquire a more ‘sophisticated’ control of the environment (Bradshaw, 1997; Fox & bin’Muhamad, 2002; Johnson-Frey, 2003). Effective tool-actions require sensing polymodal properties of (a) the agent, e.g., the effector's location and its motor properties; (b) the object, e.g. target object's location and its material properties; (c) the mean, e.g., the shape, size and functional properties of the tool. Here, we address several questions mainly related with the latter component, by investigating the effects that distinct experiences with various types of tools can produce on the multisensory representation of peri-personal space. Indeed, the sector of space surrounding the body (Rizzolatti, Fadiga, Fogassi, & Gallese, 1997) seems to be represented in primates by multisensory systems that share several functional commonalities (Làdavas, 2002; Rizzolatti, Matelli, & Pavesi, 1983).

In monkeys, multisensory processing of peri-hand space is achieved at the single cell level, as in bimodal visuo-tactile neurons that are activated both by touches delivered within the hand somatotopic receptive field (RF) and visual stimuli presented near the same RF (Bremmer, Schlack, Duhamel, Graf, & Fink, 2001; Duhamel et al., 1991, Duhamel et al., 1998; Graziano and Gross, 1995, Graziano and Gross, 1998; Rizzolatti, Luppino, & Matelli, 1998; Rizzolatti, Scandolara, Matelli, & Gentilucci, 1981). Typically, neuronal visual responses vary as a function of the distance of the visual stimulus from the hand somatosensory RF, increasing when the stimulus comes closer, and decreasing at farther distances (Duhamel, Colby, & Goldberg, 1998; Fogassi et al., 1996, Fogassi et al., 1999).

In humans, multisensory activity has been identified in possibly homologous cerebral areas by functional imaging studies (Bremmer, Schlack, Shah et al., 2001; Culham & Kanwisher, 2001; Grefkes, Weiss, Zilles, & Fink, 2002; Lloyd, Shore, Spence, & Calvert, 2003; Macaluso & Driver, 2001; Macaluso, Frith, & Driver, 2000; Weiss et al., 2000; Weiss, Marshall, Zilles, & Fink, 2003).

However, compelling evidence for functional similarities in representing peri-personal space in human and non-human primates has been provided by neuropsychological studies (di Pellegrino, Làdavas, & Farnè, 1997; Làdavas, di Pellegrino, Farnè, & Zeloni, 1998; Làdavas, Zeloni, & Farnè, 1998). In some right brain-damaged (RBD) patients with cross-modal extinction on double simultaneous stimulation (Bender, 1952; Mattingley, Driver, Beschin, & Robertson, 1997; Rapp & Hendel, 2003) contralesional tactile perception can be modulated by the distance at which ipsilesional (auditory or visual) stimuli are presented from a body-part (Farnè & Làdavas, 2002; Farnè, Demattè, & Làdavas, 2003). In the case of the hand, nearby visual stimuli (∼5 cm) are more efficient than farther ones (∼35 cm) in extinguishing contralesional tactile stimuli, this spatial modulation representing a behavioural hallmark of multisensory coding for peri-hand space (see for review Làdavas, 2002).

Because of its limited extension, peri-hand space would go little beyond the hand-reachable space when the arm is fully stretched-out. However, tools can make out-of-reach objects reachable by the hands. Furthermore, kinematics of prehensile actions performed directly by the hand or through a hand-held tool are remarkably similar (Jeannerod, 1986; Jeannerod, Arbib, Rizzolatti, & Sakata, 1995; Gentilucci, Roy, & Stefanini, 2004). The merging of sensory information from different locations (somatosensory inputs from the hand and visual inputs from the tool tip) may be useful for optimal tool-manipulation of objects that are not at hand. Indeed, multidisciplinary evidence widely supports the notion that tool-use can extend the multisensory coding of near space into far space (Làdavas & Farnè, 2004a; Maravita, Spence, & Driver, 2003; Calvert, Spence, & Stein, 2004). In a seminal work, Iriki, Tanaka, and Iwamura (1996) revealed that visual RFs of monkey's parietal neurons enlarged along the axis of a rake immediately after its use for retrieving distant food pellets. After prolonged passive tool-wielding, they also documented a backward shrinking of the same visual RFs, thus showing an activity-dependent re-mapping of far visual objects as nearer ones. Functional imaging studies have shown that the cerebral areas involved in tool-use are almost coincident with those involved in multisensory integration both in monkeys (Obayashi et al., 2001, Obayashi et al., 2002, Obayashi et al., 2003) and humans (Choi et al., 2001, Inoue et al., 2001, Moll et al., 2000; Grafton, Fadiga, Arbib, & Rizzolatti, 1997; Johnson et al., 2002; Johnson & Grafton 2003; Macaluso, Driver, & Frith, 2003).

In humans, Farnè and Làdavas (2000) reported behavioural evidence of tool incorporation in the multisensory peri-hand space by investigating cross-modal extinction in a group of RBD patients. Visual stimuli, presented at the tip of a 38 cm long rake statically held in the patients’ ipsilesional hand, induced more contralesional tactile extinction immediately after tool-use (retrieving distant objects with the rake for 5 min) than before tool-use. Stronger cross-modal extinction at the same far location after tool-use can be considered as evidence for the extension of peri-hand space along the tool axis. In the same study, backward contraction of the extended peri-hand space was also documented, as cross-modal extinction was reduced at pre-tool-use levels after a longer interval of tool inactivity. In a closely related single case study, Maravita, Husain, Clarke, and Driver (2001) similarly found that visuo-tactile extinction was stronger when the patient wielded the tip of a stick close to the visual stimulus than in absence of the stick, or when the stick was present but physically disconnected from the hand.

Several reports have now shown that tool-use can change space perception both in normal subjects (Riggio, Gawriszewski, & Umiltà, 1986; Maravita, Spence, Kennett, & Driver, 2002a; Yamamoto & Kitazawa, 2001), and neglect or extinction patients (Ackroyd, Riddoch, Humphreys, Nightingale, & Townsend, 2002; Berti & Frassinetti, 2000; Maravita, Clarke, Husain, & Driver, 2002; Pegna et al., 2001), thus raising several questions about the crucial determinants of peri-hand space extension. Is a passive change of the corporeal configuration (hand + tool) sufficient, or is some goal-directed activity needed? Is there a linear relationship between the length of a tool and the amount of peri-hand space extension? A crucial question concerns the specificity and the critical determinant of the extent to which peri-hand space increases. Does this depend upon the physical, absolute length of the tool, or the length of the tool that can be effectively used to act on objects? Here, we addressed such questions, within the same cross-modal paradigm, to shed further light onto the crucial determinants of tool dependent re-sizing of peri-hand space.

To answer the first question (passive/active experience), we investigated whether a relatively prolonged, passive exposure to a hand-held tool induces an elongation of the peri-hand space representation. In the light of the above cited neurophysiological and psychophysical findings (Iriki et al., 1996; Maravita, Spence et al., 2002; Maravita & Iriki, 2004), we expected that a passive increase in body size, physically extended by the hand-held tool, would not elongate peri-hand space representation along the tool axis.

Concerning the second question (tool-length/peri-hand space length relationships), we verified whether differently sized tools produce differential amounts of peri-hand space expansion. We predicted that, with respect to a fixed far location (60 cm from the hand), the use of a 30 cm long tool would extend peri-hand space to a much lesser degree, if any, than the use of a 60 cm long tool. Preliminary support to the first two hypotheses was also based on a single case study (Farnè, Bonifazi, & Làdavas, in press).

To answer the third question (absolute/operative length effects), we devised a hybrid tool that measured 60 cm of absolute length, but whose functionally effective part (the tines) was only 30 cm away from the hand (see Fig. 2c). We reasoned that, by dissociating within the same tool its physical aspect from its functional properties, it would be possible to demonstrate whether peri-hand space elongation is determined by the absolute length of a tool, or by its relative, functional length. In particular, if the key element is the operative length (i.e. 30 cm), then a comparable amount of peri-hand space extension should be found after use of the hybrid tool and a regular, 30 cm long tool (see Fig. 2b). Alternatively, if the absolute length (i.e. 60 cm) of the tool is crucial, then peri-hand space extension after hybrid tool-use should be similar to that obtained after the use of a regular 60 cm long tool (see Fig. 2a).

These hypotheses were tested in a group of RBD patients with left tactile extinction, who were examined in a series of conditions involving either passive exposure (1) or active use (2) of different types of tools.