Cortical control of grasp in non-human primates Thomas Brochiera,1 and Maria Alessandra Umilta`b,1 The skilled use of the hand for grasping and manipulation of objects is a fundamental feature of the primate motor system. Grasping movements involve transforming the visual information about an object into a motor command appropriate for the coordinated activation of hand and finger muscles. The cerebral cortex and its descending projections to the spinal cord are known to play a crucial role for the control of grasp. Recent studies in non-human primates have provided some striking new insights into the respective contribution of the parietal and frontal motor cortical areas to the control of grasp. Also, new approaches allowed investigating the coupling of grasp-related activity in different cortical areas for the control of the descending motor command. Addresses a Institut de Neurosciences Cognitives de la Me´diterrane´e – INCM, UMR 6193, CNRS – Universite´ de la Me´diterrane´e, 31 chemin Joseph Aiguier, 13402 Marseille 20, France b Dipartimento di Neuroscienze, Sezione di Fisiologia, Universita` di Parma, Via Volturno 39, I-43100 Parma, Italy Corresponding author: Brochier, Thomas ([email protected]
) and Umilta`, Maria Alessandra ([email protected]
Current Opinion in Neurobiology 2007, 17:637–643 This review comes from a themed issue on Motor systems Edited by Roger Lemon and Paul Bolam Available online 21st February 2008 0959-4388/$ – see front matter # 2008 Elsevier Ltd. All rights reserved. DOI 10.1016/j.conb.2007.12.002
Introduction Two main requirements are necessary to efficiently grasp an object: the capacity to transform the visual features of the object in the appropriate hand configuration and the capacity to execute and control hand and finger movements. These complex neural operations are performed by different cortical areas that are functionally and anatomically interconnected and that form a parieto-premotor neural circuit for hand grasping. Recent studies have clarified the functional properties of this cortical circuit and how it influences the corticospinal output controlling hand muscles. In the present review article we will first describe the role played by the different cortical areas involved in the control of grasp. We will emphasize that these cortical areas do not code 1
Both authors contributed equally to this work.
simple movements, but movements aimed at a specific goal, that is, goal-related motor acts like taking possession of an object . In the second part we will focus on how these areas relate to the primary motor cortex (M1) and its output to the spinal chord to control the execution of relatively independent finger movements (RIFM).
The parieto-premotor circuit for hand grasping in macaque monkeys The cortical areas forming the parieto-premotor circuit for hand grasping are the anterior intraparietal (AIP) area , the areas PF and PFG of the inferior parietal lobule (IPL), the rostral sector of the ventral premotor cortex (area F5), and the primary motor cortex (M1). Recently, the dorsal premotor area F2 was shown to be involved in the execution of visually guided grasp . This area could also be considered as part of the parieto-premotor circuit for hand grasping in the monkey. Area AIP  is located rostrally in the lateral bank of the intraparietal sulcus (IPS). The functional properties of area AIP have been extensively investigated at the single unit level [4,5,6,7] while macaque monkeys performed visually guided grasps of differently shaped 3D objects. The visual and motor responses of AIP neurons were tested in three experimental conditions: grasping in light, grasping in dark, and object observation. The results showed that some AIP neurons respond during grasping execution in light and dark, others respond only during grasping in light and finally, some neurons discharge when the monkey fixates an object even when no grasping of the object is required. There is congruence between the visual and motor responses of AIP neurons: a neuron is active when the monkey observes one particular object selectively and discharges for grasping of the same object. The observation that single neurons in AIP display a combination of visual and motor properties suggests that these neurons code the visual features of the observed objects and that together with F5 neurons (see below) they transform them into the appropriate hand configuration for grasping [8–10]. Furthermore, the reversible inactivation of AIP  produces impairment of grasping, which results in a loss of the usual match between the object shape and the appropriate preshaping of the hand. The fact that this impairment is more pronounced in the preshaping phase of grasping rather than during object manipulation emphasises the crucial role of AIP in visuomotor transformation. Current Opinion in Neurobiology 2007, 17:637–643
638 Motor systems
A series of studies showed that grasping is represented also in areas PF and PFG of the IPL [12,13,14,6]. Fogassi and co-workers showed that neurons recorded from the anterior part of IPL discharged in association with hand grasping movements. The same authors also demonstrated that most grasping neurons selectively discharge during grasp according to the final goal of the action (e.g. eating vs. placing) in which the grasp is embedded. This finding suggests that IPL neurons reflect the ‘intention’ of the agent during grasp, where the term ‘intention’ refers to the final goal of a planned action.
Area F5 is the rostral sector of the ventral premotor cortex . Area F5 is reciprocally connected with several IPL areas (PF, PFG, and AIP), with the second somatosensory area (SII), with the dorsal premotor area F2, and with the primary motor cortex (M1) [16,8,17,18]. Recently, three different architectonic fields were distinguished in F5 [19,15]: F5 convexity (F5c), F5 ‘posterior sector of posterior bank’ (F5p), F5 ‘anterior sector of posterior bank’ (F5a). These cytoarchitectonically defined sectors are strongly interconnected but are each characterized by specific patterns of connectivity with other cortical areas. The posterior bank of the inferior arcuate sulcus and the adjacent cortical convexity contain grasp-related motor neurons [8,20,21,22]. When F5 motor neurons are studied with an experimental paradigm similar to that used for AIP neurons [20,22], the results show that all tested neurons display a preference for specific objects according to the grip used to grasp them. Figure 1 shows an example of an F5 grasping neuron, whose discharge is maximal during grasping of objects differing in shape but grasped in a similar way (e.g. the cube, cone and sphere). Similarly to AIP, area F5 also contains neurons that respond to the visual presentation of objects [20,22] even when grasping is not required. The majority of these neurons (called ‘canonical neurons’), are selective both for a particular type of grip and for the observation of objects affording the same type of grip. All F5 grasping neurons fire during grasping in both light and dark. The most plausible interpretation of the visual discharge of F5 visuomotor neurons is that the visual presentation of the object activates the motor representation required to interact with the object. If a small object is observed, the hand shape to correctly interact with it, which is precision grip, is activated. Recently, grasping neurons with motor and visuomotor properties similar to those of AIP and F5 neurons have been recorded from the ventro-rostral sector of the dorsal premotor area F2, area F2vr [23,2]. Area F2vr is reciproCurrent Opinion in Neurobiology 2007, 17:637–643
One F5 motor neuron recorded while the monkey was grasping six different objects. For each object, rasters of eight trials and the resulting histogram during Movement in Light condition are presented. Small gray bars located among rasters in each trial correspond to the different events of the task. First and second horizontal lines below each histogram indicate the object presentation and object holding periods, respectively, averaged across trials. Rasters and histograms are aligned (vertical bar) with the beginning of the movement. A timescale (s) is placed on the abscissa of each histogram (from Raos et al., 2006; with permission).
Cortical control of grasp in non-human primates Brochier and Umilta` 639
cally connected with the region of area F5 (F5p) where canonical neurons are mostly located. The properties of F2vr motor and visuomotor neurons are similar to those of F5 neurons. F2vr neurons discriminate between different grips during movement execution in light and in dark. However, this selectivity is more pronounced in light, extending also to the epochs preceding movement execution (see Figure 2). These properties suggest that F2vr neurons code the continuous activation of the object representation in motor terms, but that they are more dependent than F5 neurons on the visual information during actual grasp. Taken together, the results here reviewed suggest that AIP, PF, PFG, F5p, and F2vr constitute a cortical network crucial for the visuomotor transformations of goalrelated grasping motor acts. However, at the present stage it remains unclear how the different structures of this parieto-premotor network actively interact for the control of grasp. Future experiments using simultaneous recording of neural activity from all the interconnected areas will be required to assess this topic.
Cortico-spinal control of the distal muscles for grasping We have shown that a parieto-premotor network is involved in transforming the visual properties of an object into the appropriate hand configuration for grasping [1,8]. In the following section we will investigate how this cortical network contributes to the control of the effectors
necessary for movement execution. We will review several lines of evidence supporting a hierarchical model in which area F5 influences the activity of distal muscles through cortico-cortical interactions with M1. A first line of evidence relates to the anatomical organization of the cortico-spinal and cortico-cortical projections originating from F5 and M1. F5 sends some direct projections to the spinal cord through the cortico-spinal track. However, in the macaque, the contribution of F5 to the cortico-spinal tract is rather weak. Most importantly, F5 sends very few cortico spinal projections towards the cervical enlargement where the motoneurons controlling the hand muscles are located but instead these projections terminates mainly in the upper cervical segments . An alternative pathway for F5 to influence the distal muscles is through its cortico-cortical projections to M1. The strongest input to the hand area of M1 originates from the hand representation of the premotor areas . M1 is in turn the main source of projections from the cortex toward the cervical enlargement. The cortico-motoneuronal (CM) component of these M1 projections, has a direct monosynaptic influence on the motoneurons controlling the hand and wrist muscles. Comparative studies suggest that the CM projections are essential for the powerful and precisely timed control of RIFM. Indeed, the CM projections are absent in the cat, hardly present in the squirrel monkey with a low level of manual dexterity but more developed in the most dexterous primate species such as the
Plots displaying the net normalized mean activity of the population of neurons recorded from area F2 during the execution of grasping with three different grips (best, second, and worst grips) in Movement in Light (ML) and Movement in Dark (MD) conditions. In the MD condition the ‘object presentation’ epoch corresponds to the moment in which the object was introduced, but was not visible. Grips selection was based on the net average discharge frequency in the ML condition. Resulting selection was used also for the MD condition. Depending on where the peak of the activity was, the net discharge frequency in premovement or movement epoch of the best grip in the ML condition was considered as 100. Discharge frequency of all other epochs, conditions, and grips was expressed as percentage of the peak discharge frequency. Abscissa: task epochs. Ordinate: net average discharge expressed in percent (from Raos et al. 2004, with permission). www.sciencedirect.com
Current Opinion in Neurobiology 2007, 17:637–643
640 Motor systems
macaque monkey, the great apes and the humans . In a pioneering study using retrograde transneuronal transport of rabies virus from single hand muscles in the Macaque monkey, Rathelot and Strick  demonstrated that the distribution of CM cells projecting to motoneurons of hand muscles is almost entirely restricted to M1 and area 3a (Figure 3) and does not include projections from more anterior premotor areas, such as F5 (Figure 3). These findings highlight the unique contribution of M1 to the monosynaptic control of hand and finger muscles. Secondly, the concept of a hierarchical relationship between F5 and M1 is supported by the comparison of the functional properties of these two areas in relation to hand movements. Kakei et al.  analyzed the respective contribution of populations of neurons in F5 and M1 to the control of wrist movements in different directions. These authors reported that the great majority of directionally tuned neurons in F5 displayed ‘extrinsic-like’ properties. These neurons were tuned in relation to the direction of the goaldirected movement, but their activity was not modulated by the forearm posture used for reaching. These functional properties contrasted with the intrinsic ‘muscle-like’ properties of a substantial proportion of M1 neurons that displayed a systematic shift in their preferred direction when different forearm postures were used for reaching. Some observations suggest that this distinction can be generalized to grasping movements. A recent study  shows that when an object is grasped by using a tool instead of the hand, all the F5 neurons and only half of the M1 neurons code the goal achieved by the tool and not the actual movements of the hand used to manipulate the tool. In other words for these neurons, the grasp-related activity will be similar whether the hand movement required to manipulate the tool is a flexion or an extension of the fingers. These results offer a new perspective on the functional organization of the cortical motor network for grasping that is not organized in terms of movements, but in terms of goal-directed motor acts.
Surface map of CM cells innervating digit motoneurons. The flattened maps represent the labelled CM cells in M1 following the injection of rabies virus into three digit muscles, adductor pollicis (ADP, upper panel), abductor pollicis longus (AbPL, central panel), and extensor digitorum communis (EDC, lower panel). Black arrows indicate the border between PM and M1 (areas 6/4) and between M1 and S1 (areas 4/3a). ArS, arcuate sulcus; CS, central sulcus; M, medial; R, rostral; SPcS, superior precentral sulcus (from Rathelot and Strick, 2006; with permission).
Current Opinion in Neurobiology 2007, 17:637–643
Furthermore, the functional properties of populations of neurons simultaneously recorded in F5 and M1 have been compared in a reach to grasp task toward differently shaped objects . The discharge of F5 neurons was tuned for specific grasps well before movement onset and this early tuning was carried over in the preshaping period of the task. By contrast, MI neurons lacked this early pre-movement specificity but were involved in all the movement-related periods of the task in closer relationship with the complex modulation of the activity of the hand and finger muscles during grasp. Taken together these results support the concept of a cortical circuit in which goal-directed motor acts are represented in the activity of F5 neurons and are translated into motor www.sciencedirect.com
Cortical control of grasp in non-human primates Brochier and Umilta` 641
commands that are appropriate for movement execution in M1 . A final line of evidence supporting the hierarchical organization of the F5-M1 system derives from the observation that the activity of hand-related neurons in M1 is modulated by the cortical input from F5. Tokuno and Nambu  reported that in the anesthetized monkey, pyramidal tract neurons in M1 respond at short latency to the intracortical stimulation of the premotor cortex. These authors suggested that these cortico-cortical interactions are of particular importance for the preparatory control of forelimb movements. This hypothesis is supported by the experiment of Cerri et al.  showing that the stimulation of the ventral premotor area F5 can produce a robust facilitation of the EMG response of intrinsic hand muscle to the single-shock stimulation of M1 (Figure 4). Similar facilitation effects are also observed in the response evoked by the stimulation of M1 in arm and particularly hand motoneurons . One can therefore speculate that these interactions play an essential role for the visual to motor transformation along the premotor– motor pathway for the control of the hand and digit muscles during grasp.
Conclusions The present review highlighted the importance of the cerebral cortex in the control of grasp. We have reviewed the contribution of multiple areas in the parietal and frontal lobes to the visuomotor transformation related to hand grasping. We have also showed that when the goal of grasping is experimentally dissociated from the hand movements required for actual grasp, this goal is specifically represented in the activity of all F5 neurons but only by some M1 neurons. This finding sheds a new light on the basic organization through which the cortical motor system controls hand grasping. This organization appears to be based not only on the control of grasping movements, but also on the coding of the goals of forthcoming actions.
Facilitation of M1 outputs by stimulation of the F5 portion of PMv. Panels (a–d) show the average EMG responses of the thenar muscle to the intracortical stimulation of the contralateral cortex (50 sweeps/average). (a) Lack of EMG response to a double-shock stimulation of F5 (2 mA 70 mA, 3-ms separation). (b) EMG response to a single shock stimulation of M1 (1 mA 70 mA); the 23 superimposed unrectified sweeps shown below indicate the large variability of the EMG evoked response. (c) Facilitation of the responses evoked from M1 by the conditioning stimulation of F5 (condition-test interval: 3 ms). The consistency of the evoked response is seen on the 23 superimposed
In addition, we explored how this parieto-premotor activity relates to the descending cortico-spinal outputs to the motor nuclei of the spinal cord. We observed that M1 makes a unique contribution to the control of hand and finger movements through its direct monosynaptic CM projections. This led us to discuss a hierarchical model in which F5 exerts an essential modulation of M1 activity for the control of the distal muscles during grasp.
Acknowledgements MA Umilta` is supported by grants from EU projects Neurobotics (Contract IST/FET 001917) and NeuroProbes (Contract IST-027017) and MIUR (Ministero Italiano dell’Universita` e della Ricerca). sweeps below the average trace. (d) Subtraction (average in c minus average in b) shows the additional effect of the conditioning shocks (from Cerri et al., 2003, with permission). Current Opinion in Neurobiology 2007, 17:637–643
642 Motor systems
References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: of special interest of outstanding interest 1.
Rizzolatti G, Fogassi L, Gallese V: Cortical mechanisms subserving object grasping and action recognition: a new view on the cortical motor functions. In The Cognitive Neurosciences, edn. 2. Edited by Gazzaniga MS. MIT Press; 2000: 539-552.
Raos V, Umilta` MA, Gallese V, Fogassi L: Functional properties of grasping-related neurons in the dorsal premotor area F2 of the macaque monkey. J Neurophysiol 2004, 92:1990-2002.
Luppino G, Murata A, Govoni P, Matelli M: Largely segregated parietofrontal connections linking rostral intraparietal cortex (areas AIP andVIP) and the ventral premotor cortex (areas F5 and F4). Exp Brain Res 1999, 128:181-187.
Sakata H, Taira M, Murata A, Mine S: Neural mechanisms of visual guidance of hand action in the parietal cortex of the monkey. Cereb Cortex 1995, 5:429-438.
Murata A, Gallese V, Luppino G, Kaseda M, Sakata H: Selectivity for the shape, size, and orientation of objects for grasping in neurons of monkey parietal area AIP. J Neurophysiol 2000, 83:2580-2601.
Gardner EP, Babu KS, Reitzen SD, Ghosh S, Brown AS, Chen J, Hall AL, Herzlinger MD, Kohlenstein JB, Ro JY: Neurophysiology of prehension. I. Posterior parietal cortex and object-oriented hand behaviors. J Neurophysiol 2007, 97(1):387-406.
Gardner EP, Babu KS, Ghosh S, Sherwood A, Chen J: Neurophysiology of prehension: III. Representation of object features in posterior parietal cortex of the macaque monkey. J Neurophysiol 2007, 98(2):1656-1670.
Rizzolatti G, Luppino G: The cortical motor system. Neuron 2001, 31:889-901.
Jeannerod M, Arbib MA, Rizzolatti G, Sakata H: Grasping objects: the cortical mechanisms of visuomotor transformation. Trends Neurosci 1995, 18:314-320.
10. Fagg AH, Arbib MA: Modeling parietal-premotor interactions in primate control of grasping. Neural Netw 1998, 11:1277-1303. 11. Gallese V, Murata A, Kaseda M, Niki N, Sakata H: Deficit of hand preshaping after muscimol injection in monkey parietal cortex. Neuroreport 1994, 5:1525-1529. 12. Yokochi H, Tanaka M, Kumashiro M, Iriki A: Inferior parietal somatosensory neurons coding face-hand coordination in Japanese macaques. Somatosens Mot Res 2003, 20:115-125. 13. Fogassi L, Ferrari PF, Gesierich B, Rozzi S, Chersi F, Rizzolatti G: Parietal lobe: from action organization to intention understanding. Science 2005, 308:662-667. 14. Fogassi L, Luppino G: Motor functions of the parietal lobe. Curr Opin Neurobiol 2005, 15:626-631. Interesting and updated review focused on some recent monkey data on the role of the PPC in visuomotor transformation for grasping. This work also presents recent data showing that the PPC is crucially involved in motor cognitive functions, such as coding of motor intention, action and intention understanding, and construction of peripersonal visuomotor space representations. 15. Belmalih A, Borra E, Contini M, Gerbella M, Rozzi S, Luppino G: A multiarchitectonic approach for the definition of functionally distinct areas and domains in the monkey frontal lobe. J Anat 2007, 211:199-211. 16. Geyer S, Matelli M, Luppino G, Zilles K: Functional neuroanatomy of the primate isocortical motor system. Anat Embryol 2000, 202:443-474. 17. Dum RP, Strick PL: Frontal lobe inputs to the digit representations of the motor areas on the lateral surface of the hemisphere. J Neurosci 2005, 25:1375-1386. 18. Borra E, Belmalih A, Calzavara R, Gerbella M, Murata A, Rozzi S, Luppino G: Cortical connections of the macaque anterior Current Opinion in Neurobiology 2007, 17:637–643
intraparietal (AIP) area. Cereb Cortex (August)2007. [epub ahead of print] PMID: 17720686. 19. Nelissen K, Luppino G, Vanduffel W, Rizzolatti G, Orban GA: Observing others: multiple action representation in the frontal lobe. Science 2005, 310:332-336. First correlation between cytoarchitectonically defined cortical regions of interest and functional properties assessed by fMRI in macaques monkey. 20. Raos V, Umilta` MA, Murata A, Fogassi L, Gallese V: Functional properties of grasping-related neurons in the ventral premotor Area F5 of the macaque monkey. J Neurophysiol 2006, 95:709729. This paper describes in depth the properties of hand grasping neurons recorded in area F5. The neurons have been recorded with the same paradigm that has been used to study the properties of hand grasping neurons in the dorsal premotor area F2, and in the anterior intraparietale area (AIP), a comparison of the functional properties of the three cortical areas (F5, F2, AIP) is discussed for the first time. 21. Hoshi E, Tanji J: Distinctions between dorsal and ventral premotor areas: anatomical connectivity and functional properties. Curr Opin Neurobiol 2007, 17:234-242. Exhaustive review on the functional properties and the reciprocal role of PMd and PMd in many aspects of motor behavior that include motor preparation, action planning, action selection and action execution. The authors also describe the anatomical connectivity that characterizes each of the two premotor cortical regions. 22. Umilta` MA, Brochier T, Spinks RL, Lemon RN: Simultaneous recording of macaque premotor and primary motor cortex neuronal populations reveals different functional contributions to visuomotor grasp. J Neurophysiol 2007, 98:488-501. This work, for the first time, address the issue of the relative contributions of primary motor cortex (M1) and area F5 to visually guided grasp of different objects by simultaneous multiple electrode recordings from the hand representations of these two areas in macaque monkeys. The authors, by comparing the visual and motor properties, the time course of the response and the grip specificity of the two populations of neurons show the specific role played by areas M1 and F5 during grasp. 23. Raos V, Franchi G, Gallese V, Fogassi L: Somatotopic organization of the lateral part of area F2 (dorsal premotor cortex) of the macaque monkey. J Neurophysiol 2003, 89:15031518. 24. He SQ, Dum RP, Strick PL: Topographic organization of corticospinal projections from the frontal lobe: motor areas on the lateral surface of the hemisphere. J Neurosci 1993, 13:952980. 25. Nakajima K, Maier MA, Kirkwood PA, Lemon RN: Striking differences in transmission of corticospinal excitation to upper limb motoneurons in two primate species. J Neurophysiol 2000, 84:698-709. 26. Rathelot JA, Strick PL: Muscle representation in the macaque motor cortex: an anatomical perspective. Proc Natl Acad Sci U S A 2006, 103:8257-8262. In this important study using retrograde labelling of CM cells with rabies virus, the authors analyze the representation of hand muscles in the primary motor cortex of macaque monkeys. They report that the representation of each individual muscle is widespread and overlap extensively with the representation of other muscles. These observations have important implications for the functional organization of the corticospinal control of grasping movements. 27. Kakei S, Hoffman DS, Strick PL: Direction of action is represented in the ventral premotor cortex. Nat Neurosci 2001, 4:1020-1025. 28. Umilta` MA, Escola L, Intskirveli I, Grammont F, Rochat M, Caruana F, Jezzini A, Gallese V, Rizzolatti G: When pliers become fingers in the monkey motor system. PNAS 2008, Jan 31 [epub ahead of print] PMID: 18238904. The authors address the issue of goal representation in the motor cortex by studying the properties of cortical motor neurons (recorded from the premotor area F5 and the primary motor cortex) during motor acts that have the same final goal (grasping) but require opposite hand movements to achieve it. The results showed that when the goal to grasp is dissociated from the movements required to accomplish it, part of the primary motor cortex controls the movement execution while ventral premotor area F5 seems to code the goal of the action. www.sciencedirect.com
Cortical control of grasp in non-human primates Brochier and Umilta` 643
29. Brochier T, Spinks RL, Umilta` MA, Lemon RN: Patterns of muscle activity underlying object specific grasp by the macaque monkey. J Neurophysiol 2004, 92:1770-1782.
31. Cerri G, Shimazu H, Maier MA, Lemon RN: Facilitation from ventral premotor cortex of primary motor cortex outputs to macaque hand muscles. J Neurophysiol 2003, 90:832-842.
30. Tokuno H, Nambu A: Organization of nonprimary motor cortical inputs on pyramidal and nonpyramidal tract neurons of primary motor cortex: an electrophysiological study in the macaque monkey. Cereb Cortex 2000, 10:58-68.
32. Shimazu H, Maier MA, Cerri G, Kirkwood PA, Lemon RN: Macaque ventral premotor cortex exerts powerful facilitation of motor cortex outputs to upper limb motoneurons. J Neurosci 2004, 24:1200-1211.
Current Opinion in Neurobiology 2007, 17:637–643