Different left brain regions are essential for grasping a tool compared with its subsequent use

Abstract

Tool use engages a left hemispheric network including frontal, temporal and parietal regions. Patients with left brain lesions (LBD patients) exhibit deficits when demonstrating use of a single tool (apraxia). When attempting to use a tool, some apraxic patients show errors in the preceding grasping movement.

Forty-two LBD patients and 18 healthy controls grasped individual tools and demonstrated their typical use. For patients with a tool use impairment (22), lesion analysis revealed a large area of overlap in the left hemisphere, mainly in the supramarginal gyrus (SMG). For patients with erroneous grasping (12), the lesion overlay showed overlaps in the left frontal and parietal cortices, especially in the inferior frontal gyrus (IFG) and the angular gyrus (ANG). However, contrasting lesions associated with impaired grasping versus tool use impairments reveal little overlap, limited to the inferior parietal cortex. Presumably the left IFG is involved in selection processes in the context of tool use, such as choosing a functional or non-functional grasping movement depending on the task and the online information about the tool's structure and orientation. The ANG might provide this grasp related information, which is relevant for the specific action.

The contribution of the SMG to tool use involves more general principals, such as integrating online and learned tool use information into the action plan for the use movement.

Introduction

Everyday we manipulate different recipient objects with different kinds of tools. For example we regularly use a toothbrush to clean our teeth, or we use a knife for slicing bread. Brain damage caused by stroke can lead to apraxia, an impairment involving for example defective motor actions during tool use pantomime (e.g. pantomiming the movement of hitting a nail with a hammer), demonstration of the use of single tools (e.g. demonstrating the movement only with the hammer), or actual tool use with a tool and a recipient (e.g. hammer and nail). Apraxic errors include searching for the right movement, producing only partially correct motions (for example trying to press a nail with a hammer into a piece of wood instead of using a hitting movement), substituting other tool use movements (for example stirring with a hammer) or even complete omissions (Goldenberg, 2008, Poeck, 1997). Most patients who exhibit tool use pantomime deficits, improve when they are allowed to actually use the tool (Goldenberg and Hagmann, 1998, Goldenberg et al., 2004). Apraxia is particularly associated with lesions in the left hemisphere. Accordingly, neuroimaging studies of tool use actions indicate that a left-lateralized hemispheric network is responsible for carrying out tool-related tasks. This network includes parietal, frontal and temporal regions (Beauchamp and Martin, 2007, Frey, 2008, Hermsdörfer et al., 2007, Lewis, 2006, Martin, 2007, Weisberg et al., 2007). It is assumed that relatively independent cognitive components in this distributed network have to be retrieved and integrated for the planning of complex tool use actions, such as manipulation knowledge and functional knowledge, (Assmus et al., 2007, Grossman et al., 2002). Depending on its localization within this network, a lesion could cause different forms of difficulties concerning tool handling. Buxbaum and Saffran (2002) inferred from their lesion study that knowledge about tool manipulation is located in the left fronto-parietal cortex. Tool manipulation involves the production of an appropriate grip posture on a tool's handle. However, impaired knowledge about the purpose of a tool was related to damages in the temporal lobe (Hodges et al., 2000). Further, the ability to infer a tool's function directly from its structural properties is also important for the use of novel tools; or the use of familiar tools in an alternative way. This ability can compensate for defective tool use knowledge and also seems to depend on the function of the left parietal cortex (Spatt et al., 2002, Hodges et al., 2000, Goldenberg and Hagmann, 1998). In a study with 38 left brain damaged patients, Goldenberg and Spatt (2009) showed that inferior frontal lesions are related to the impaired retrieval of functional associations, as well as to the impaired use of common and novel tools. However, parietal lesions were specifically related to impairment in selection and application of common tools (supramarginal gyrus) and novel tools (dorsal inferior parietal cortex and supramarginal gyrus), though not to impairment in retrieval of functional knowledge. Summarized, tool use planning seems to be dependent on temporal, frontal and parietal regions of the left hemisphere. But it is remarkably often reported that the left parietal cortex is involved in different kinds of tool use actions. A common component of familiar tool use could be the integration of perceived artefact-properties and tool use knowledge into an action plan. It is assumed that the integration of tool use knowledge and sensory-motor information is carried out by the left parietal cortex (Buxbaum et al., 2007, Frey, 2007).

Like many object related actions, tool use actions typically start with a reaching and grasping movement. The important features for prehensile movements, like hand shaping and arm transport, have to be preplanned (Gentilucci et al., 1997, Jeannerod, 1984, Klatzky et al., 1995, Rosenbaum et al., 1990, Spatt et al., 2002). Preplanning includes the anticipation of the properties of an object to be grasped and takes relevant features of the subsequent action into account (Armbrüster and Spijkers, 2006, Marteniuk et al., 1987, Short and Cauraugh, 1997, Zhang and Rosenbaum, 2008). In particular, when a tool has to be grasped, semantics¹ seem to influence the type of grasping. Creem and Proffitt (2001) investigated healthy participants grasping tools which were oriented with their handle pointing away from the subject's body. The participants’ task was to grasp and transport the presented tool. In 72% of the trials subjects rotated their hand into a pronated posture with the thumb directed towards the functional part of the tool (functional grasping). If the action goal is to use a tool, a rotated grasp will facilitate the subsequent use movement. That means a biomechanically more awkward grasp is preferred to benefit a subsequent movement phase. Additionally Creem and Proffitt (2001) showed, with a dual-task design, that when grasping is performed simultaneously with the recall of paired associates, the number of appropriate grasps dropped to 17%. The authors concluded that the (unperturbed) retrieval of semantic information is necessary to grasp a tool according to its functional identity. But components other than semantics (in the sense of knowledge about how to apply a tool) seem to influence grasping as well. Recently we studied the grasping behavior of stroke patients, who suffered from damage to the left (LBD) or right hemisphere (RBD) respectively and healthy controls with the Creem & Proffit - Paradigm (Randerath et al., 2009). A significant effect of subsequent task type and of apraxia on the preceding grasping movement was found when the tool was presented with its handle pointing away from the participant. When participants had to pick up the tool to use it, functional grasping was produced more frequently (nearly 100% for all non-apraxic participants) than when they had to pick up the tool to transport it into a container. This supported Creem & Proffit's interpretation. Retrieval of tool use information (semantics) necessary for the use task might elicit functional grasping. We concluded that, in addition to the functional information about a tool, the structural information (division in handle and functional part) also contributes to the selection of the appropriate grasp type. The following noticeable findings argued for this reasoning. During the use task, inappropriate non-functional grasping occurred rarely and was exclusive to the LBD group with apraxia. Indeed non-functional grasping was always followed by erroneous tool use, but errors in the using action appeared more frequently, thus also after correct functional grasping. Therefore the perception of the tool's structure could have automatically led to functional grasping, making the grasping action more robust compared to more semantic dependent use action. The behavior of apraxic patient (I.L.) was noticeable in that she produced more functional grasping during the transport-task than during the use task. These findings suggested that an interference effect appeared in the use task, because both information types (functional semantics and structural information about the presented tool) had to be integrated in parallel into the movement plan before the grasping movement started. Preserved processing of structural information may support correct grasping in many apraxic patients despite impaired functional knowledge.

The neural correlates for grasping a tool in an efficient way to use it subsequently have not been reported yet. In studies with monkeys it was shown that the inferior parietal cortex (IPL) and the ventral premotor area (F5) are involved in planning grasping movements (Fogassi and Luppino, 2005, Fogassi et al., 2005, Rizzolatti and Matelli, 2003). Borra et al. (2008), showed that the dorsal (intraparietal region) and ventral stream (superior temporal sulcus and medial temporal gyrus) as well as frontal regions (F5) are interconnected in the monkey's brain. It is assumed that the dense interconnection between the IPL and area F5 supports the transformation and integration of object information while planning a specific grasping posture. The function of IPL is said to be the integration of knowledge stored in the ventral stream with sensory information processed by the dorsal stream (Borra et al., 2008, Fogassi and Luppino, 2005, Rizzolatti and Matelli, 2003).

The posterior parietal cortex, including the inferior and the superior parietal cortex, are thought to be analogically organized in humans. While there is not an area F5 defined in a human brain, the inferior frontal gyrus (IFG) shows similarities concerning the localization and the cytoarchitecture (Rizzolatti and Arbib, 1998). The findings in the literature clearly indicate that in humans IFG and IPL are similarly involved in object and grasp related tasks (Grafton et al., 1997, Grezes et al., 2003, Tunik et al., 2008a). The use of transcranial magnetic stimulation (TMS) on IFG and IPL in a study by Tunik et al. (2008a), with 10 healthy subjects, caused a delay in goal-directed hand–object interactions (e.g. reach to grasp a cup and turn it over as if to pour liquid into it). In accordance with these findings, Grezes et al. (2003) showed that IFG and IPL are activated when objects (cylinders) have to be grasped. More dorsal regions, like the intraparietal sulcus and the superior parietal cortex (SPL) seem to play an additional role in grasping. During grasping the anterior IPS is associated with processing object properties (e.g. length and orientation of rectangular objects) (Culham et al., 2003, Tunik et al., 2005). The SPL is said to specifically be involved in the online adaptation of grasping movements. Effects have been shown for the left SPL, when affordances are manipulated while the subject is executing a grasping movement (e.g. changing the object size) (TMS: Glover et al., 2005; EEG: Tunik et al., 2008b).

In summary, the IPL and the IFG seem to be involved in object grasping. When tools have to be used according to their function several cortical regions of the left hemisphere, but particularly the parietal lobe, seem to be engaged. One could assume that the reported left temporo–parieto–frontal network is not only responsible for planning tool use actions, but also for planning related grasping actions. In a fMRI study by Creem-Regehr et al. (2007), the authors found similar activations in the left premotor, parietal and temporal regions, when participants imagined grasping tools, or when they imagined grasping objects.

The aim of this study is to find out which regions are specifically important for grasping tools functionally for utilization. In addition, the above-described findings raise the question of whether there is an overlap between the regions responsible for functional grasping and those associated with execution of tool use actions. Based on the results of these studies, which either investigated functional application of tools or grasping objects, we predict, that different left brain regions are essential for grasping a tool compared with its subsequent use. While the left IFG might contribute to planning a grasping movement, the left IPL is expected to be engaged in both grasp planning and tool use.

To clarify this point, we used our formerly applied paradigm (Randerath et al., 2009) and analyzed the lesion data of 42 stroke patients with left brain damage (LBD). A lesion overlap for impaired demonstration of tool use and for the production of inappropriate non-functional grasping is expected to be located in the left IPL.