Anatomical and spatial matching in imitation: Evidence from left and right brain-damaged patients.

Neuropsychologia ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia

Anatomical and spatial matching in imitation: Evidence from left and right brain-damaged patients Paola Mengotti a,n, Enrico Ripamonti b, Valentina Pesavento c, Raffaella Ida Rumiati a a

Neuroscience Area, SISSA, Trieste, Italy Department of Economics, Management and Statistics, Statistical Section, University of Milan-Bicocca, Milan, Italy c S.C. Medicina Riabilitativa, Azienda Ospedaliero-Universitaria ‘Ospedali Riuniti’, Trieste, Italy b

art ic l e i nf o

a b s t r a c t

Article history: Received 23 December 2014 Received in revised form 24 June 2015 Accepted 29 June 2015

Imitation is a sensorimotor process whereby the visual information present in the model's movement has to be coupled with the activation of the motor system in the observer. This also implies that greater the similarity between the seen and the produced movement, the easier it will be to execute the movement, a process also known as ideomotor compatibility. Two components can influence the degree of similarity between two movements: the anatomical and the spatial component. The anatomical component is present when the model and imitator move the same body part (e.g., the right hand) while the spatial component is present when the movement of the model and that of the imitator occur at the same spatial position. Imitation can be achieved by relying on both components, but typically the model's and imitator's movements are matched either anatomically or spatially. The aim of this study was to ascertain the contribution of the left and right hemisphere to the imitation accomplished either with anatomical or spatial matching (or with both). Patients with unilateral left and right brain damage performed an ideomotor task and a gesture imitation task. Lesions in the left and right hemispheres gave rise to different performance deficits. Patients with lesions in the left hemisphere showed impaired imitation when anatomical matching was required, and patients with lesions in the right hemisphere showed impaired imitation when spatial matching was required. Lesion analysis further revealed a differential involvement of left and right hemispheric regions, such as the parietal opercula, in supporting imitation in the ideomotor task. Similarly, gesture imitation seemed to rely on different regions in the left and right hemisphere, such as parietal regions in the left hemisphere and premotor, somatosensory and subcortical regions in the right hemisphere. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Ideomotor compatibility Apraxia Correspondence problem Right hemisphere Gesture imitation

1. Introduction One conceptualization of imitation considers it as an act of copying someone's movements. As such imitation requires an interaction between at least two actors: the model, i.e., the person executing the movement in the first place, and the imitator, i.e., the person copying the movement. To accomplish imitation, the imitator needs to integrate the sensory information coming from the visual system with the motor system in order to reproduce the movement. How this integration is accomplished is known as the correspondence problem (Brass and Heyes, 2005) and it is still matter of debate. It is therefore critical to investigate how the observed actions are mapped onto the motor system of the observer. n Correspondence to: Cognitive Neuroscience, Institute of Neuroscience and Medicine (INM-3), Research Centre Jülich, Leo-Brandt-Str. 5, 52425 Jülich, Germany. E-mail address: [email protected] (P. Mengotti).

The production of a simple action has been shown to be facilitated or interfered by the simple concurring vision of a similar or different movement, suggesting that imitation is indeed based on sensorimotor interaction. In one of these studies (Brass et al., 2000), participants were required to move one of two fingers in response to a spatial cue (i.e., a cross placed on the hand stimulus on the screen) while observing a moving hand as task-irrelevant cue. Results showed that when the movement performed by the participants was the same as the one performed by the hand stimulus, their reaction times were smaller than in the opposite condition, that is when the two movements differed. Subsequent studies manipulated this basic paradigm (Brass et al., 2001a,, 2003; Bertenthal et al., 2006; Longo et al., 2008; Longo and Bertenthal, 2009; Boyer et al., 2012) and replicated the original observation that action is modulated by perception. According to this view, the integration between perception and action is achieved through a process of common coding or ideomotor compatibility (Prinz, 1997; Brass et al., 2000, 2001a; Hommel et al., 2001; Massen and Prinz, 2009) between the model and the imitator. This occurs because

http://dx.doi.org/10.1016/j.neuropsychologia.2015.06.038 0028-3932/& 2015 Elsevier Ltd. All rights reserved.

Please cite this article as: Mengotti, P., et al., Anatomical and spatial matching in imitation: Evidence from left and right brain-damaged patients. Neuropsychologia (2015), http://dx.doi.org/10.1016/j.neuropsychologia.2015.06.038i

2

P. Mengotti et al. / Neuropsychologia ∎ (∎∎∎∎) ∎∎∎–∎∎∎

percepts and action plans share common properties and it is this similarity that allows imitation to be achieved: the more similar is the perceived movement with the movement to be produced, the easier is the production of the movement (Prinz, 1997; Massen and Prinz, 2009). Moreover, the degree of similarity or ideomotor compatibility between imitator's and model's movements can be based on two different parameters: the anatomy of the model and the location in space of the model's movements. The anatomical imitation is based on the anatomical matching with the model; for instance, if the model moves his/her left arm, the imitator will move his/her left arm, with movements being performed in different positions of the space. The “mirror” or spatial imitation consists in replicating the movement as if the imitator was in front of a mirror. Thus when the model moves his/her left arm, the imitator will move his/her right arm, to spatially match the model. As the anatomical and spatial components have been teased apart in several studies (Bertenthal et al., 2006; Mengotti et al., 2012; Mengotti et al., 2013b), it is likely that the two processes rely on different cognitive processes (Boyer et al., 2012; but see also Heyes and Ray, 2004 for a different account). The basic matching mechanisms of imitation have been studied using paradigms of ideomotor compatibility involving very simple finger movements (Brass et al., 2000, 2001a, 2003; Bertenthal et al., 2006; Longo et al., 2008; Longo and Bertenthal, 2009; Boyer et al., 2012; Mengotti et al., 2012, 2013b). However, more often imitation is used, for instance, to learn more complex gestures. After brain damage, this ability to imitate gestures can be impaired; this disorder is known as ideomotor apraxia (see Goldenberg (2009) for an historical review). Patients with ideomotor apraxia show a deficit in imitating actions and/or performing them on verbal command and these deficits cannot be attributed to elementary motor and sensory deficits, aphasia, agnosia or frontal inertia (De Renzi and Faglioni, 1999; Cubelli et al., 2000; see Rumiati et al., 2010, for a review). According to the classical model proposed by Liepmann (1920), action control is achieved in two steps: the generation of the mental image of the intended gesture and the implementation of this mental image into the appropriate motor output. A failure occurring at the first step will give rise to ideational apraxia, clinically characterized by the inability to use objects, whereas a failure at the second step will give rise to ideomotor apraxia. Cognitive neuropsychological models provided a coherent conceptualization of how imitation comes about, by proposing that the behavior depends on the nature of the input and the output involved in a given task (Rothi et al., 1991; Cubelli et al., 2000; Tessari and Rumiati, 2004). According to these models, imitation is suggested to be accomplished by relying on two main pathways: the semantic pathway, which encompasses the semantic systems and is used for reproducing known gestures, and the direct pathway, used for meaningless gestures, which bypasses the semantic system and allows a direct reproduction of the visual input into motor output. These models allowed generating predictions as to how imitation can break down after brain damage, with selective deficits depending on the particular component of the model that is damaged (see, Rumiati et al., 2010 for a review). Therefore, patients with different lesions will show a deficit in imitation of meaningless or meaningful gestures (Tessari et al., 2007). Further dissociations are observed in imitation performance depending on the part of the body that it is involved, with lesions differentially affecting the ability to imitate hand postures or finger movements (Goldenberg, 1999). The predominant role of the left hemisphere in supporting the ability to imitate gestures is widely recognized (Liepmann, 1920; De Renzi et al., 1980; Papagno et al., 1993; Goldenberg, 1995; Haaland et al., 2000; Tessari et al.,

2007); nonetheless the right hemisphere seems to contribute to imitation (Goldenberg and Karnath, 2006; Goldenberg et al., 2009), in particular when the visuo-spatial analysis of the movement is more important, as for the imitation of meaningless gestures (Tessari et al., 2007; Rumiati et al., 2005). Moreover, right brain-damaged patients' imitation performance is more impaired with meaningless gestures (Tessari et al., 2007) and finger configurations (Goldenberg, 1999, 2009; Della Sala et al., 2006), whereas left brain-damaged patients' imitation performance is more impaired with hand postures (Goldenberg, 1999), suggesting a division of labor between the two hemispheres. In the present study, we aimed at better understanding the matching processes that sustain imitation by applying a paradigm based on the ideomotor compatibility. This paradigm is particularly useful because it allows studying the effect on patients' performance of the anatomical or spatial matching between the model's and imitator's movements. This has been investigated when both matching processes were present or when the two movements matched at the anatomical or at the spatial level. Participants were asked to reproduce a tapping movement performed by the model in two different ways: in the anatomical subtask, they were instructed to move the finger that matched the model's finger based on the anatomical identity, whereas in the spatial subtask, participants were instructed to move the finger that matched the model's finger based on its location in space. When the model was presented in a mirror perspective, the imitator's and the model's movement matched both for the anatomical identity of the body part moved and for their spatial location in space, whereas when the model was presented in a non-mirror perspective the two movements matched only for one of the two features, either in their anatomical identity or in their spatial location. Moreover, we analyzed patients' performance on a more complex gesture imitation task, in which they reproduced intransitive gestures performed by a model in a spatial (i.e., mirror) perspective or in an anatomical perspective. This task is a standardized test usually adopted in the neuropsychological assessment to detect deficits in imitation (Tessari et al., 2015). Indeed, tasks similar to our ideomotor task have been used in neuroimaging studies (Iacoboni et al., 1999; Brass et al., 2001b; Koski et al., 2003; Bien et al., 2009; Mengotti et al., 2012) while gesture imitation tasks are more common in neuropsychological studies (De Renzi et al., 1980; Papagno et al., 1993; Goldenberg, 1995; Haaland et al., 2000; Tessari et al., 2007; Mengotti et al., 2013a). These two lines of research led to different results about the localization of the cognitive processes underlying imitation. Neuroimaging studies with healthy individuals showed consistently bilateral activations of premotor and frontal regions and activations of the parietal operculum (Iacoboni et al., 1999; Brass et al., 2001b; Koski et al., 2003; Bien et al., 2009; Mengotti et al., 2012), whereas in neuropsychological studies lesions to left parietal regions have been more consistently associated with deficits in imitation of gestures (Haaland et al., 2000; Weiss et al., 2001; Tessari et al., 2007; Mengotti et al., 2013a). Only a few studies in which gesture imitation was investigated using neuroimaging with healthy participants (Rumiati et al., 2005; Menz et al., 2009) found bilateral fronto-parietal activations, thus suggesting also some a contribution of the right hemisphere to imitation. However, no study directly compared ideomotor and gesture imitation tasks in a sample of brain-damaged patients. Previous evidence allowed scholars to argue in favor of the existence of two distinct mechanisms that can be used to solve the corresponding problem in imitation: a process of spatial compatibility that supports spatial imitation and a process of matching of body parts of model and imitator that supports anatomical imitation. In other words, a mechanism based on the spatial (or



mirror) correspondence, and another based on the anatomical correspondence with the model (Bertenthal et al., 2006; Boyer et al., 2012; Mengotti et al., 2012, 2013b). We analyzed the performance of left and right brain-damaged patients in order to investigate whether signatures of these two matching processes can be found both in the ideomotor task and in a more complex gesture imitation task. Moreover, we investigated which brain networks underpin such different matching mechanisms. We hypothesize that the performance of left and right braindamaged patients would be selectively impaired when imitation is based only on the spatial or anatomical matching with the model, therefore we expect a predominant role of the left hemisphere in processing the anatomical component of imitation, and a predominant role of the right hemisphere in processing the spatial component.

2. Methods 2.1. Participants All participants confirmed their voluntary participation by signing the informed consent and all were able to understand task instructions at the time of testing. The study was in accordance with the ethical standards of the local Ethics Committee and in accordance with the Declaration of Helsinki of 1975, as revised in 1983.

3

Table 1 Patients' demographical data. Case

Sex

Age

Education (y)

Testing post-onset (m)

LBD AR CF GD GP PC PG PP RE SS TS

M M M M F M F F M M

58 80 67 49 75 63 45 40 50 65

5 17 – 17 8 8 8 8 12 8

2 2 3.5 6.5 1 1.5 3 2.5 1.5 1

RBD BE CF CP DBC GE GL HB MB MR PM

M M F F M F F F M M

64 65 71 38 61 79 68 62 42 53

5 11 11 8 8 5 4 8 5 11

5.5 3 1 10 21 3 2 2 2 1.5

Abbreviations y¼ years; m¼ months; M¼ male; F ¼ female.

2.2. Experimental study 2.1.1. Patients A total of 20 patients (8 females; mean age 60713 years, and mean education 97 4 years) took part in the study. They were recruited from a number of patients consecutively admitted to the rehabilitation units of the Ospedali Riuniti in Trieste and of the San Camillo neurorehabilitation hospital in Venice. Only patients with focal left- or right- unilateral brain lesions and no previous neurological history were included. They had all normal or correctedto-normal vision, no hearing difficulties, at least 5 years of education and they were not older than 80. Ten patients had left hemisphere brain damage (LBD) and 10 patients had right hemisphere brain damage (RBD); the groups of left and right brain-damaged patients were matched for age (t (18) ¼0.19) and education (t(18) ¼ 1.6). See Table 1 for demographical data. All patients underwent a neuropsychological evaluation, to assess general intelligence, language functions, executive functions, memory, visuo-spatial and attentional abilities (see Table 2). 2.1.2. Controls A total of 20 healthy individuals (10 females; mean age: 63 710 years; education: 10 73 years) participated in the study. They were matched with the group of patients for age (t(38) ¼ 0.8) and education (t(38) ¼ 1.4). All healthy controls had no previous neurological history and did not have any neurological dysfunction. The Mini Mental State Examination (Folstein et al., 1975) was administered to ensure that control participants were not suffering from any form of cognitive decline (mean score: 27, range 24– 30). Half of the controls performed the experimental tasks using their left hand in order to allow their performance to be compared with that of the LBD patients (who used their unimpaired left hand for performing the task), and the other half performed the experimental tasks with their right hand in order to be compared with the RBD patients (who used their unimpaired right hand for performing the task). Only one healthy participant was considered ambidextrous, while all others were right-handed when handedness was evaluated using the Edinburgh Handedness Inventory (Oldfield, 1971).

The tasks were administered to the patients in two different sessions in the same week, whereas the controls performed all the tasks in one single session. The order of the tasks was randomized across sessions and participants. Prior to the experimental tasks, patients performed a simple left-right judgment task, to assure that they were able to perform the left-right discrimination. No deficits in performing this task were detected. All patients and controls completed the following tasks (see also Fig. 1). 2.2.1. Ideomotor task On each trial, five-frame video sequences were presented on a black background of a 15.6″ computer screen, depicting a single tapping movement that could be performed by either a left (50% of the trials) or a right hand with either the index (50% of the trials) or the ring finger. In each trial, the first frame was presented for 1000 ms, followed by three frames for 40 ms each, depicting the intermediate positions of the finger, and a final frame depicting the end position for 2000 ms. Each trial was followed by a 3000 ms inter-trial interval (black screen). E-Prime 2 software (Psychology Software Tools, Pittsburgh, PA) was used for stimulus presentation and data collection. Reaction times (RTs) and accuracy data were collected, RTs being measured from the first of the three 40 ms frames that followed the presentation of the first 1000 ms frame. Participants were asked to reproduce tapping movements similar to those performed by the hand stimulus, by pushing two different buttons, one with the index finger and one with the ring finger. They performed two different subtasks, the anatomical subtask and the spatial subtask, presented in two different blocks, with 48 trials per block. The order of the blocks was counterbalanced across participants. In the anatomical subtask, they were asked to move the finger of the assigned hand (left or right, depending on the group) that was anatomically compatible with the finger moving on the screen. For instance, when the video displayed an index finger movement, they had to tap their index finger, irrespective of whether the observed hand was a left or a right hand, and of whether the moved finger occupied the same or


4

LBD

Language

Case EHI AAT token AR CF GD GP PC PG PP RE SS TS

100 100 100 100 100 100 100 60 100 100

29 9 25 35 36 14 37 1 11

STM

AAT repetition 84 134 47 86 – 128 103 132 141 129

AAT written

AAT naming

34 46 38 35 – 64 65 60 84 42

86 109 94 50 – 54 115 75 87 95

AAT compreh. 97 115 85 96 – 99 110 86 97 90

LTM

Attention

Visuo-perceptual

Praxis

Picture naming

FAS

Span forward

Span Corsi LTM backward faces

Rey figure

Words list (IR)

Words list TMT (DR) A

TMT B

Matrices Weigl Raven's CPM

VOSP scr.

VOSP obj.

BORB MFM

IA IMA

– – – – 27 – – – – –

– – – – 15.6a – 15.1 – – 4.2

– 5 – – 4.25 4.25 3 5 4.75 4.25

– – – – 2 – – t.i. 5 3

– – – 19.25 t.i. – 13.75 15 – –

– 34.6 – – 57.3 – – – – –

– 10.8 – – 9.3 – – – – –

– t.i. – t.i. – 308 127 343 – –

37.75 – – 20 – 36.25 – – 32 42.25

19 20 – – – – – 20 – –

18 18 – – 16 17 – 18 – 18

– – – 25 – – – – – –

14 7 14 – 14 – 14 10 14 14

4.75 5.25 – 4.5 3.25 4.25 4 4 3.5 –

23 – – – – 18 – – – 23

Language

– 35 – 73 170 29 53 97 – –

RBD

Neglect

Case EHI

BIT tot.

BIT line cross.

BIT line bisect.

BIT star canc.

BIT letter canc.

BIT copy

BIT draw.

Picture naming

FAS

Words list (DR)

TMT A

TMT B

BE CF CP DBC GE GL HB MB MR PM

76 59 139 132 129 94 128 132 132 134

36 18 36 36 36 36 34 36 35 36

5 0 7 7 7 7 9 4 6 2

18 22 52 53 47 31 46 50 54 51

12 14 37 33 38 16 33 36 35 38

3 3 4 2 1 2 3 3 0 4

2 2 3 1 0 2 3 3 2 3

29 30 30 30 30 – 26 – 30 29

20.9 24 31.9 31 32 – 43.9 36.5 27.4 22.5

– 6.3 11.8 – – – 6.1 12 10.2 8.1

– – 57 62 75 – 43 102 – –

– – 436 147 133 – 382 287 – –

r 100 r/l 100 100 r 100 100 r 90

Executivelogical

Attention

7.25 11 – 9.75 9.25 4.75 8.75 1.75 8 2.75

31.8 27.5 – 31.1 20.3 36 32.8 25.3 35.3 28

Executive-logical

70 39 68 62 62 66 68 72 65 70

Visuo-perceptual

Praxis

Matrices Weigl Raven's CPM

VOSP scr.

VOSP obj.

IA

IMA

– – 50 50.75 42.75 25.25 – 46.25 – 33

19 – – 19 20 – – – – 19

17 – – – – – – – – 16

– – 14 – – – – – – 14

53 56 62 69 64 62 68 60 70 66

13.75 8.75 12 – – – 13.75 8.5 – 11

27.9 – 25.6 21.4 27.3 17 25.9 15.4 27.3 30.3

Tests used in the neuropsychological assessment of the LBD patients: EHI¼ Edinburgh Handedness Inventory (Oldfield, 1971); AAT ¼ Aachener Aphasie Test (Luzzatti et al., 1996); AAT token ¼ subtest for comprehension; AAT repetition ¼subtest for repetition; AAT written ¼ subtests for reading and writing; AAT naming ¼ subtest for production; AAT compreh. ¼ subtests for auditory and visual comprehension; Picture naming ¼in-house developed test for language production, maximal score ¼ 30; FAS ¼ phonemic fluency test (Carlesimo et al., 1996); Span forward¼ Digit Span Forward for verbal short-term memory (Orsini et al., 1987); Span backward ¼ Digit Span Backward for shortterm memory and working memory; Corsi ¼Corsi test for spatial short-term memory (Spinnler and Tognoni, 1987); LTM faces ¼long-term memory, face recognition (Warrington, 1996); Rey figure¼delayed recall of the Rey– Osterrieth Complex Figure (Caffarra et al., 2002); Words list ¼recall of a list of words not semantically related (Mauri et al. 1997); IR¼ immediate recall; DR ¼ delayed recall; TMT A ¼ Trail Making Test for attention; TMT B ¼ Trail Making Test for executive functions (Giovagnoli et al., 1996); Matrices ¼test of the matrices for attention (Spinnler and Tognoni, 1987); Weigl ¼Weigl's tests for executive functions (Spinnler and Tognoni, 1987); Raven's CPM ¼Raven's Coloured Progressive Matrices for general intelligence (Carlesimo et al., 1996); VOSP ¼Visual Object and Space Perception battery (Warrington and James, 1991). VOSP scr.¼ screening task; VOSP obj. ¼ object decision task; BORB MFM ¼ Minimal Feature Match Test of the Birmingham Object Recognition Battery (Riddoch and Humphreys, 1993); IA¼ test for ideational apraxia (De Renzi and Lucchelli, 1988); IMA ¼ test for ideomotor apraxia (Tessari et al., 2015). t.i.¼ test interrupted because the patient was not able to perform the task. Patients are sorted alphabetically by their initials. Tests used in the neuropsychological assessment of the RBD patients: EHI ¼ Edinburgh Handedness Inventory (Oldfield, 1971); r¼ test not available, patient report of being right-handed; r/l ¼ test not available, patient report of being ambidextrous; BIT ¼ behavioral inattention test (Wilson et al., 1987); BIT tot.¼ BIT total score; BIT line cross. ¼ subtest of line crossing; BIT line bisect.¼ subtest of line bisection; BIT star canc. ¼subtest of star cancellation; BIT letter canc. ¼ subtest of letter cancellation; BIT copy ¼subtest of picture copying; BIT draw. ¼ subtest of drawing; Picture naming ¼ in-house developed test for language production, maximal score ¼ 30; FAS ¼ phonemic fluency test (Carlesimo et al., 1996); Words list ¼recall of a list of words not semantically related (Mauri et al., 1997); DR ¼delayed recall; TMT A ¼Trail Making Test for attention; TMT B ¼ Trail Making Test for executive functions (Giovagnoli et al., 1996); Matrices¼ test of the matrices for attention (Spinnler and Tognoni, 1987); Weigl ¼ Weigl's tests for executive functions (Spinnler and Tognoni, 1987); Raven's CPM ¼ Raven's Coloured Progressive Matrices for general intelligence (Carlesimo et al., 1996); VOSP ¼Visual Object and Space Perception battery (Warrington and James, 1991). VOSP scr.¼ screening task; VOSP obj. ¼object decision task; IA¼ test for ideational apraxia (De Renzi and Lucchelli, 1988); IMA ¼test for ideomotor apraxia (Tessari et al., 2015). Patients are sorted alphabetically by their initials. a

Capasso and Miceli (2001).



Table 2 Patents' neuropsychological assessment. Scores have been already corrected for age and education when necessary. Scores in bold are those that were below normal range.


5

Fig. 1. Experimental tasks. The two main experimental tasks are schematized. Left: ideomotor task: (i) the anatomical subtask, whereby patients were asked to move the finger that matched anatomically the one moved by the model; (ii) the spatial subtask, whereby patients were asked to move the finger that matched the spatial position of the one moved by the model. Showing left and right hands as models led to a compatible condition with the presence of both the anatomical and spatial matching between the model's and imitator's movement, and to a non-compatible condition with the presence only of the anatomical or spatial matching with the model. In both tasks the hand pictures represent the model, whereas the hand sketches represent the imitator. Right: The gesture imitation was performed in two different conditions: (i) the spatial condition, in which the imitator is asked to mirror the movement of the model, by moving the arm in the same spatial location (i.e., the model moves the left arm and the imitator the right arm); (ii) the anatomical condition, in which the imitator is asked to move the same arm used by the model (i.e., the left arm is moved both by the model and the imitator). In both conditions the model is presented in the upper part of the figure, whereas the imitator is the one showing her back.

a different position in the space, relative to the observed moving finger. In the spatial subtask, they had to move the finger of their assigned hand that was spatially compatible with the finger moving on the screen. For instance, when the observed finger movement occurred closer to the right side of the screen (as in case of a right hand tapping with its index finger), they had to tap using the finger closest to the same side, irrespective of its anatomical identity (i.e., the ring finger when performing the task with their right hand). Using both left and right hands as stimuli, two conditions were generated. First, the compatible condition, in which the hand stimulus was a mirror-image of the participants' hand, with finger movements being matched both anatomically and spatially (i.e., movements were made on the same side of the space and with the same finger). In this condition the same response was required both in the anatomical and in the spatial subtask. Second, the noncompatible condition, in which the non-mirror hand was shown. In this condition different responses were required in the anatomical and in the spatial subtask. In the anatomical subtask, the finger movement performed by the participant matched that of the model only anatomically; both the model and the participant, for instance, moved the index finger, but movements occurred in different locations of the space. In the spatial subtask, the finger movement shown by the model matched that of the participant only spatially, for example both the model and the participant moved the finger on the left side of the screen, hence the finger moved by the model and the participant were different. The noncompatible condition in the anatomical subtask was the only condition in which the spatial matching between the stimulus and the participants' hand did not occur, whereas the non-compatible condition in the spatial subtask was the only condition in which the anatomical matching between the stimulus and the participants' hand did not occur. As the effect in the non-compatible condition is embedded in the opposite subtask, an impairment in the non-compatible condition is suggestive of an increased difficulty in reproducing the movement due to the decrease in the similarity between participant's and model's gestures, and not to an explicit difficulty in performing the task. Indeed in the compatible condition the movements match both at the anatomical and at the spatial level, whereas in the non-compatible condition

they match only at the spatial level in the anatomical task and at the anatomical level in the spatial task. On the contrary, an impairment in both compatible and non-compatible conditions in one of the two subtasks suggests a difficulty in performing imitation following the explicit instructions of the task, e.g., the ability to imitate on the basis of the spatial position of the movements. 2.2.2. Gesture imitation task Patients and controls were asked to imitate, one after the other, 18 meaningful intransitive gestures and 18 meaningless gestures, derived from the meaningful ones by modifying the spatial relationship between the effector and the main body axis (Tessari et al., 2015). The presentation of the meaningful and meaningless gestures was blocked and the meaningless gestures were presented always first, to prevent participants from selecting the direct, non-semantic, route for both types of gestures (Tessari and Rumiati, 2004). Gestures were always repeated twice. The test was presented to the participants in two different conditions. In the anatomical condition, the experimenter performed the gestures with the arm that was anatomically compatible with the participants' arm. In this case, if the experimenter showed the gestures with his/her right arm, the participant performed the task with his/her right arm. In the spatial condition, the experimenter performed the gestures with the arm that was spatially compatible with the participants' arm. In this case, if the experimenter showed the gestures with his/her left arm, in a mirror-like way, the participant performed the task with his/her right arm. For each item a score of 0, 1, 2 was given according to the performance (0 ¼ no correct imitation, 1¼ correct imitation in the one of the two trials, 2 ¼ correct imitation in both trials), with the cut-off varying according to age and education, for a total score of 72 maximum. Participants' performance was videotaped and analyzed offline by three independent raters, who evaluated each trial as “correct” or “incorrect”. If at least two out of three raters assessed the performance as “correct”, 1 point was assigned to that trial, otherwise the trial was considered as “incorrect” and was assigned 0 points. The patients' performance in the spatial condition is reported in Table 2, as the procedure used in this condition overlaps with the standard testing for ideomotor apraxia (Tessari et al., 2015).



6

2.3. Behavioral analysis

2.5. Lesions analysis

Analyses of accuracy were performed on the individual percentage of correct responses in each experimental condition. RTs analyses were performed on the median scores for each experimental condition; error trials and trials in which RTs were above two standard deviations from the mean RTs for that condition were excluded. Following this criterion, in the ideomotor task the mean number of trials included in the analysis for each condition for the patients was 18.7 (on average across all patients), with a range from 16.9 to 20.9 between conditions (24 trials per condition were presented). Since patients' and controls' RTs were not normally distributed, Wilcoxon signed rank test and Mann– Whitney U test were used to compare performance within and across (patients vs. controls; LBD vs. RBD patients) groups, respectively. Spearman's nonparametric correlations coefficients were calculated to measure the strength of association between scores in the experimental tasks and demographic/neuropsychological variables. In order to detect single-patients' deficits with respect to the control group in each experimental condition, individual t scores were computed using the software released with the article by Crawford and Garthwaite (2005). Data for each participant were entered as percentage of correct responses. The software provides a t score for each individual performance and estimates the abnormality of the individual score with respect to the mean performance of the correspondent control sample.

Computed tomography (CT) or magnetic resonance imaging (MRI) scans were available for all patients. For lesion mapping, the CT/MRI scan that was the closest in time to the neuropsychological evaluation was chosen. Lesions were mapped by two independent neuropsychologists (P.M. and E.R., both uninformed as to the patients' behavioral performance) from the axial CT/MRI scans into the standard Montreal Neurological Institute (MNI) template (Holmes et al., 1998) adopted by the MRIcro software (Rorden and Brett, 2000; available online http://www.mccauslandcenter.sc.edu/mricro/in dex.html). The inclination of the scan on the orbito-meatal plane and the correspondence between the orbito-meatal plane and the bicommissural plane inclination were kept into account. The template was rotated manually (Z dimension only) to match the approximate slice plan of the patients' scan using gyri, sulci and deep gray matter structures as anatomical landmarks. Each region of interest (ROI) can be rotated back into the standard MNI space to allow group comparisons. Finally, the ROIs traced in MRIcro are transformed in VOIs for the behavioral analysis with MRIcroN (Chris Rorden, available online: http://www.mccauslandcenter.sc. edu/mricro/mricron/index.html). Gyri and white matter tracts within ROIs have been identified using MRIcroN and the Anatomy Toolbox (Eickhoff et al., 2005) of the Statistical Parametrical Mapping software (SPM, Wellcome Department of Cognitive Neurology, http://www.fil.ion.ucl.ac.uk/spm) implemented in the MATLAB (2010b, version 7.11, The Mathworks Inc., http://www. mathworks.com) environment. We first inspected our sample by overlaying the lesions of each group of patients, thus creating map of the regions more often lesioned in the group of the left brain-damaged patients, and another map of the regions more often lesioned in the group of the right brain-damaged patients. We then analyzed our lesion data adopting the subtraction approach (Rorden and Karnath, 2004). Using MRIcroN, the lesions of all patients with defective performance, on the one hand, and those of all patients with normal performance on the other hand, were overlapped separately for each condition. Subsequently, the overlap of the control group was subtracted from the overlap of the patients that presented a deficit in the condition of interest. The lesion map generated by the subtraction highlights only the regions that are functionally involved in the task, independently of the anatomical vulnerability of the regions. Patients' performance was considered defective or within normal range based on individual t scores, calculated as described in Section 2.3. This procedure was performed for each condition of the ideomotor and of the gesture imitation tasks that had, at least, two patients performing pathologically. As subtractions were performed between groups of different sizes, relative percentages rather than absolute values were used. Only those regions in which the difference of percentage of overlapping lesions for a specific group of patients (after subtraction from the control group) are above 55% have been shown in Figs. 7 and 8.

2.4. Error analysis The qualitative analysis of errors in gesture imitation was performed, based on criteria used in previous studies (Tessari et al., 2007; Carmo and Rumiati, 2009) and slightly modified to adapt to the present test. Error types include: (i) Spatial error of the hand: the overall movement of the limb is correct, but the hand posture is wrong; (ii) Spatial error of the arm: the movement is recognizable, but it is performed with the arm forming a wrong angle with the body; (iii) Spatial error of the fingers: the movement is recognizable, both the position of the arm and the hand are correct, but the finger posture is wrong; (iv) Mislocation: the overall movement is correct, but the movement endpoint is not correctly reproduced; (v) Semantic errors, which are further divided into two subcategories: (a) Prototypicalization: participants reproduce the prototypical version of the meaningful action instead of the one presented. (b) Visual-semantic: an action visually similar and semantically related to the target action is produced. (vi) Visual: an action visually similar to the target-action is produced. Visual errors can be further divided into three subcategories: (a) Perseveration: it involves the repetition of an action, or part of an action, that has been previously presented (note that it was not possible to distinguish between motor and visual perseveration). (b) Lexicalization: a meaningful action, visually similar to the meaningless target action but not included in the list, is produced; (c) Substitution: a visually similar meaningful action, not included in the list, is produced instead of the meaningful action that was presented. (vii) Omission: the imitation of the target action is omitted; (viii) Unrecognizable gesture: the response involves a movement that the raters failed to recognize.

3. Results 3.1. Behavioral analysis The performance of the two control groups, one using the right hand and the other using the left hand to perform the tasks, did not significantly differ in all conditions of the ideomotor task and gesture imitation task (Mann–Whitney U tests, p 40.6). Patients' and controls' performance for accuracy and RTs is shown in Table 3.



7

Table 3 Group performance across all experimental tasks. Accuracy scores (ACC) are shown as percentage of correct responses and reaction times (RTs) are expressed in milliseconds. Standard error of the mean is shown in brackets. Ideomotor task

Gesture imitation task

Anatomical Compatible

LBD Controls RBD Controls

Spatial Non-compatible

Compatible

ACC

RTs

ACC

RTs

ACC

96.3 (.02) 98.3 (.01) 82.5 (.06) 97.5 (.02)

936 (75) 778 (72) 980 (118) 650 (75)

79.6 (.07) 98.3 (.01) 80.8 (.08) 97.1 (.01)

1135 (84) 893 (76) 882 (110) 737 (83)

93.8 (.04) 98.3 (.01) 80.9 (.07) 99.6 (.004)

3.1.1. Accuracy We first considered the ideomotor task. We compared the distribution of accuracy responses within each group. LBD patients showed a difference between compatible and non-compatible conditions (Wilcoxon, p ¼0.008, level of significance was set at 0.025, Bonferroni corrected threshold) in the anatomical subtask, with lower performance in the non-compatible condition, i.e., the condition in which the spatial matching between model and participants' hand is not present (96% of correct responses in the compatible condition vs. 80% of correct responses in the noncompatible condition). No difference in performance between conditions or subtasks was found for RBD patients and for the control groups. We compared then accuracy scores between patients and control groups. LBD patients showed impairment with respect to controls only in the non-compatible condition in the anatomical subtask (80% of correct responses for the LBD patients vs. 98% of correct responses for controls; Mann–Whitney, p ¼0.002; level of significance was set at 0.0125, Bonferroni corrected threshold), and not in the compatible condition (Mann–Whitney, p¼ 0.53). Moreover, LBD patients showed a performance within normal range in both conditions of the spatial subtask (Mann–Whitney, compatible p ¼0.48, non-compatible p¼ 0.44). Performance of RBD

Anatomical

Spatial

Non-compatible RTs

735 662 1029 525

(92) (78) (197) (52)

ACC

RTs

ACC

ACC

93.8 (.04) 99.6 (.004) 80.2 (.06) 98.7 (.01)

744 (83) 646 (57) 926 (205) 484 (48)

83.6 (.06) 96.4 (.01) 81.9 (.04) 95.1 (.02)

89.1 (.04) 97.1 (.01) 87.5 (.02) 96.9 (.01)

patients was below the normal range both in compatible and noncompatible conditions of the spatial subtask (for the compatible condition: 81% of accuracy for RBD patients vs. 99.6% of accuracy for controls; for the non-compatible condition: 80% of accuracy for RBD patients vs. 99% of accuracy for controls; Mann–Whitney, compatible p o0.0001, non-compatible p¼ 0.002; level of significance was set at 0.0125, Bonferroni corrected threshold), whereas their performance in the anatomical task in both compatible and non-compatible conditions was not different from that of controls (Mann–Whitney, compatible p¼ 0.03, non-compatible p¼ 0.02; level of significance was set at 0.0125, Bonferroni corrected threshold). Furthermore we compared performance between patients' groups. RBD patients showed a performance in the non-compatible condition in the spatial subtask that was below that of LBD patients, with a trend towards significance (Mann–Whitney, p¼ 0.019; level of significance was set at 0.0125, Bonferroni corrected threshold). All other conditions were not different between RBD and LBD patients (Mann–Whitney, spatial compatible p¼ 0.029, anatomical compatible p ¼0.052, anatomical non-compatible p ¼0.85). Fig. 2 shows the accuracy scores of all groups in the ideomotor task.

Fig. 2. Accuracy scores for the ideomotor task. Accuracy scores for LBD and RBD patients for anatomical and spatial subtask in ideomotor task are shown. Dotted lines indicate the score for the respective control group. Asterisks indicate significant differences between patients' and controls' scores or between conditions (n ¼p o 0.0125).


8


We then considered the gesture imitation task. We compared the distribution of accuracy responses within each group. LBD patients performed the gesture imitation task better in the spatial condition than in the anatomical condition (84% of accuracy in the anatomical condition vs. 89% of accuracy in the spatial condition; Wilcoxon, p o0.005; level of significance was set at 0.025, Bonferroni corrected threshold). No difference in performance between conditions was found for RBD patients and for the control groups. Moreover we compared accuracy scores between patients and control groups. LBD patients were more impaired than matched controls in the anatomical condition (84% of accuracy for LBD patients vs. 96% of accuracy for controls; Mann–Whitney, p ¼0.004, level of significance was set at 0.025, Bonferroni corrected threshold), however they showed a performance in the spatial condition that was within normal range (Mann–Whitney, p ¼0.03, level of significance was set at 0.025, Bonferroni corrected threshold). RBD patients showed impaired performance compared to controls in both anatomical and spatial conditions (for the anatomical condition: 82% for RBD patients vs. 95% for controls; for the spatial condition: 87% for RBD patients vs. 97% for controls; Mann–Whitney, p ¼0.004 and p ¼0.002, level of significance was set at 0.025, Bonferroni corrected threshold). No differences were found between LBD and RBD patients for gesture imitation. We performed an additional analysis in order to investigate patients' imitation of meaningful and meaningless gestures, as the gesture imitation task included both types of gestures. We compared the distribution of accuracy responses within each group. RBD patients performed the gesture imitation task better with meaningful than meaningless gestures, in both the anatomical and spatial conditions (Wilcoxon, p ¼0.005 and p o0.005 respectively; level of significance was set at 0.0125, Bonferroni corrected threshold). No difference in performance between conditions was found for LBD patients and for the control groups. Moreover we compared accuracy scores between patients and control groups. LBD patients were more impaired than matched controls in imitating meaningless gestures in the anatomical condition (Mann– Whitney, p ¼0.002, level of significance was set at 0.0125, Bonferroni corrected threshold), however their performance in imitating meaningless gestures in the spatial condition was within normal range (Mann–Whitney, p ¼0.04, level of significance was set at 0.0125, Bonferroni corrected threshold). RBD patients showed impaired performance compared to controls when asked to imitate meaningless gestures in both anatomical and spatial conditions (Mann–Whitney, p ¼0.003 and p o0.001, level of significance was set at 0.0125, Bonferroni corrected threshold). No differences were found between LBD and RBD patients when comparing imitation of meaningful and meaningless gestures in the two conditions. Fig. 3 shows the accuracy scores of all groups in the gesture imitation task. Taken together these results suggest that imitation (in both the ideomotor and gesture imitation tasks) is easier for LBD patients than RBD patients when patient's and model's movements are matched spatially. However, the same perspective seems to be detrimental for RBD patients' performance in either imitative tasks. As for RBD patients, correlations between experimental tasks did not show a possible effect of neglect on performance in the ideomotor task. A positive correlation was only found between the anatomical condition of the gesture imitation task and the Behavioral Inattention Test (BIT; N ¼10, Spearman's R ¼0.78, p ¼0.008), suggesting an effect of neglect on patients' performance in this particular condition. Indeed, in this anatomical condition, gestures were presented in the left side of the patient's visual space. Therefore, we further investigated the influence of neglect on the RBD patients' anatomical imitation, by looking only at patients

Fig. 3. Accuracy scores for the gesture imitation task. Accuracy scores for LBD and RBD patients for anatomical and spatial conditions in gesture imitation are shown. Dotted lines indicate the score for the respective control group. Asterisks indicate significant differences between patients' and controls' scores or between conditions (n ¼ p o 0.025).

without neglect (n ¼6). Indeed the anatomical imitation improved consistently at the group level, from 82% of accurate responses in the complete group (n ¼10) to 88% (only non-neglect patients, n¼ 6), whereas for spatial imitation there was only a slight increase, from 88% of accurate responses in the complete group (n ¼10) to 90% (only non-neglect patients, n ¼6). Single-patients' performance in the ideomotor and gesture imitation tasks is shown in Table 4. 3.1.2. Reaction times We investigated RTs differences in the ideomotor task between patients and healthy controls. We first compared the compatible and non-compatible conditions within each of the two subtasks, to control for performance differences when the anatomical or the spatial matching with the model was or was not present. LBD patients (199 ms of compatibility effect; Wilcoxon, p ¼0.005; level of significance was set at 0.0125, Bonferroni corrected threshold) and both control groups presented an effect of the spatial compatibility with the model (116 ms and 87 ms of compatibility effect; Wilcoxon, p ¼0.005 and p¼ 0.01, for the LBD control group and for the RBD control group respectively; level of significance was set at 0.0125, Bonferroni corrected threshold), showing a significant difference between compatible and non-compatible conditions in the anatomical subtask, with faster reaction times in the compatible condition than in the non-compatible condition. RBD patients did not show this difference (Wilcoxon, p ¼0.96). LBD patients did not show differences in RTs with respect to controls in both tasks. RBD patients showed slower performance than controls in both compatible and non-compatible conditions in the spatial subtask (for the compatible condition: 1029 ms for the RBD patients vs. 525 ms for the controls; for the non-compatible condition: 926 ms for the RBD patients vs. 484 ms for the controls; Mann–Whitney, p¼ 0.003 and p¼ 0.009, respectively; level of significance was set at 0.0125, Bonferroni corrected threshold). The performance in both compatible and non-compatible condition of the anatomical subtask was within normal range (Mann–Whitney, p ¼0.029 and p ¼0.19; level of significance was set at 0.0125, Bonferroni corrected threshold). No RTs differences were found when comparing LBD with RBD patients. Fig. 4 shows the RTs of all groups in the ideomotor task.



9

Table 4 Individual performance of all patients across all experimental tasks. Scores are shown as percentage (%) of correct responses and as t scores (t). t scores in bold are those that were below normal range. Ideomotor task

Gesture imitation task

Anatomical Case

Compatible

Spatial Non-compatible

%

t

Anatomical

Compatible %

t

Non-compatible

%

t

LBD AR CF GD GP PC PG PP RE SS TS

100 96 100 96 100 100 100 79 92 100

0.40 0.55 0.40 0.55 0.40 0.40 0.40 4.60 1.50 0.40

88 83 100 92 92 79 96 21 83 63

4.91 7.29 0.81 3.00 3.00 9.20 1.10 36.85 7.29 16.83

96 100 100 100 71 100 75 100 96 100

0.55 0.40 0.40 0.40 6.51 0.40 5.55 0.40 0.55 0.40

100 96 100 100 75 100 67 100 100 100

RBD BE CF CP DBC GE GL HB MB MR PM

100 75 100 79 92 75 100 75 42 88

0.30 2.68 0.30 2.18 0.69 2.68 0.30 2.68 6.65 1.19

87 17 100 83 96 63 100 79 92 92

2.29 19.16 0.69 3.29 0.31 8.25 0.69 4.27 1.29 1.29

96 79 79 42 100 92 96 38 92 96

2.79 14.96 14.96 42.47 0.29 5.79 2.79 44.81 5.79 2.79

96 79 71 46 88 46 96 89 100 92

3.2. Error analysis As for gesture imitation, patients of both groups made more errors in the anatomical condition than in the spatial condition (see Table 5). Nevertheless, a qualitative analysis of gesture imitation showed a different pattern of results in LBD and RBD patients. In particular, LBD patients committed more hand than finger spatial errors, characterized by a wrong hand posture, whereas RBD patients committed more finger than hand spatial errors, characterized by wrong finger posture or wrong finger selection. Irrespective of the type of condition, LBD patients committed 56% and RBD patients

Spatial

%

t

%

t

%

t

0.30 2.64 1.00 0.96 18.04 0.30 23.91 0.30 0.30 0.30

94 36 88 85 79 89 92 100 89 85

0.57 14.4 2 2.72 4.15 1.76 1.05 0.86 1.76 2.72

97 54 94 86 86 92 94 100 90 97

0.03 13.7 0.98 3.53 3.53 1.62 0.98 0.92 2.26 0.03

0.69 4.65 6.65 12.61 2.67 12.61 0.69 2.19 0.31 1.67

71 51 86 92 81 79 88 90 85 97

4.42 8.09 1.67 0.57 2.58 2.95 1.3 0.93 1.85 0.35

74 78 86 96 89 86 94 83 97 92

7.28 6.01 3.46 0.29 2.51 3.46 0.92 4.42 0.03 1.56

30% of hand spatial errors (out of the total amount of errors). In contrast, RBD and LBD patients committed 41% and 16% finger spatial errors respectively. Other type of errors, such as spatial error of the arm and mislocation are generally equi-distributed across the two groups (spatial errors of the arm: 16% for LBD and 13% for RBD; mislocations: 7% for LBD and 8% for RBD). Error distribution for LBD and RBD patients is shown in Fig. 5. 3.3. Lesions analysis Lesions of all patients for each group, i.e., LBD and RBD, were overlapped in order to show the lesions' distribution for each

Fig. 4. Reaction times for the ideomotor task. Reaction times (in ms) for LBD and RBD patients for anatomical and spatial subtasks in ideomotor task are shown. Dotted lines indicate the score for the respective control group. Asterisks indicate significant differences between patients' and controls' scores or between conditions (n ¼ p o0.0125).



10

Table 5 Distribution of types of errors across groups and experimental conditions for gesture imitation. Errors are presented as percentage relative to the total amount of errors. LBD

Spatial error hand Spatial error arm Spatial error finger Mislocation Prototypicalization Visual-semantic Perseveration Lexicalization Substitution Omission Unrecognizable gesture Total no. errors

RBD

Anatomical

Spatial

Anatomical

Spatial

47.9 18.5 16.8 7.6 – – 4.2 1.7 0.8 – 2.5 119

64.1 12.8 15.4 6.4 – – – – – – – 78

24.8 18.8 39.1 9 – – 4.5 2.3 – – 0.8 133

36.3 7.7 44 7.7 – – 4.4 – – – – 91

Fig. 5. Error types in the gesture imitation task. Distribution of error types is presented (shown in percentage of the total amount of errors) for LBD and RBD patients with respect to anatomical and spatial conditions in the gesture imitation task.

group and the regions that were more frequently lesioned. Both groups showed a more frequent involvement of the superior temporal gyrus, insular cortex, inferior frontal gyrus, putamen, and fronto-parietal operculum (see Table 6 and Fig. 6). Moreover, RBD and LDB patients did not differ for lesion size (t(18) ¼ 0.19). In the subtraction analysis we considered only those conditions in which there were at least two patients of each group, i.e., impaired vs. unimpaired, as established on the basis of the t scores calculated on individual patients' performance and shown in Table 4 (the compatible condition for the anatomical ideomotor subtask for the LBD patients and the compatible condition for the spatial ideomotor subtask for the RBD patients were not considered in the subtractions, as they did not meet the criterion). In the following analyses, the critical regions that, when lesioned, gave rise to a selective deficit associated with a particular condition were obtained by subtracting the lesioned regions of patients without a deficit (i.e., control group) from the lesioned regions of patients with a deficit in the condition of interest. We first considered the ideomotor task. For LBD patients, regions lesioned in patients with impaired imitation in the noncompatible condition of the anatomical subtask, after subtracting the control group, were manly located in the parietal operculum

(both the OP1 and OP2, see Eickhoff et al., 2007), the superior temporal gyrus, as well as subcortical structures as the putamen, the superior longitudinal fasciculus and the external capsule (see part A of Fig. 7). As for the spatial subtask, the subtraction analysis showed that the inferior frontal gyrus, including Broca's area (BA 44) was the region that, when lesioned, led to impairment in both compatible and non-compatible conditions. Moreover, the compatible condition was also associated with lesions to the superior longitudinal fasciculus, whereas the non-compatible condition was also associated with lesions to the superior temporal gyrus (BA 22, see part B of Fig. 7). For RBD patients, in the anatomical subtask, the regions damaged specifically in patients with impaired performance in the compatible condition, after subtracting the patients without defective performance, were the inferior frontal gyrus (BA 45), the parietal operculum (OP4 and OP3), the superior temporal gyrus and, as regards the white matter, the superior longitudinal fasciculus, the fornix and the retrolenticular portion of the internal capsule. Regions specifically damaged in patients with impaired performance in the non-compatible condition were the precentral gyrus, involving the primary motor (BA 4) and somatosensory (BA 3) cortices, the parietal operculum (OP3 and OP1), the supramarginal gyrus (BA 41), the thalamus and, with respect to the white matter, the superior longitudinal fasciculus, the superior part of the corona radiata and the retrolenticular portion of the internal capsule (see part C of Fig. 7). As for the spatial subtask, the regions lesioned in patients with impaired performance in the non-compatible condition were the parietal operculum (OP3 and OP4), the superior temporal and the insular cortices, and the superior longitudinal fasciculus (see part D of Fig. 7). As for gesture imitation in LBD patients, in both anatomical and spatial conditions, the subtraction analysis showed that the regions that, when lesioned, led to impaired imitation were the angular (BA 39) and supramarginal gyri (BA 40), the superior and middle temporal gyri, the superior and middle occipital gyri, part of the superior longitudinal fascicle, and part of the superior and posterior portion of the corona radiata. In addition, lesions of the premotor cortex and inferior frontal gyrus were associated with lower performance only in the anatomical condition (see part A of Fig. 8). RBD patients showed a distribution of lesioned regions, obtained in the subtraction analysis, which was very different from that of the LBD patients, involving in both the anatomical and the spatial conditions, the precentral and postcentral gyri, including the primary motor (BA 4), the somatosensory cortices (BA 3), the premotor cortex (BA 6), the superior temporal gyrus and the putamen. Moreover, RBD patients presented a large involvement of the white matter, including the superior longitudinal fasciculus, the anterior and superior part of the corona radiata, the external capsule and the internal capsule. In the anatomical condition further regions were recruited, such as the inferior frontal gyrus (BA 44), the insular cortex and the globus pallidus (see part B of Fig. 8).

4. Discussion In the present study we analyzed the behavior of patients with unilateral left or right focal brain-damage, who performed an ideomotor task and a gesture imitation task in order to investigate the effects of the anatomical and spatial matching between the model and the imitator. In line with our predictions, we found important differences between LBD and RBD patients depending on the presence or the absence of the anatomical or spatial matching with the model. RBD patients showed a deficit in imitation when they relied on the spatial matching with the model. This is particularly evident when



11

Table 6 Description of lesion for all patients. Case LBD AR

CF GD

GP

PC

PG

PP

RE SS

TS

RBD BE

CF

CP DBC

GE

GL

HB

Regions affected by the lesion

Precentral gyrus, frontal operculum, insula, parietal operculum, postcentral gyrus, Heschl gyrus, superior temporal gyrus, parietal inferior, supramarginal and angular gyri. Putamen. Rentrolenticular part of internal capsule, superior part corona radiata, external capsule, superior longitudinal fasciculus. Superior and middle temporal gyri, parietal inferior, supramarginal and angular gyri, middle occipital gyrus. Superior longitudinal fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus, superior temporal gyrus. Caudate, putamen. Anterior limb internal capsule, anterior and superior part corona radiata, external capsule, superior longitudinal fasciculus, superior fronto-occipital fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, postcentral gyrus, middle and inferior temporal gyri, parietal superior and inferior, supramarginal and angular gyri, cuneus, superior, middle and inferior occipital gyri. Putamen, pallidum. Superior, rentrolenticular part of internal capsule, posterior part of corona radiata, posterior thalamic radiation, sagittal stratum, external capsule, superior longitudinal fasciculus. Precentral gyrus, frontal operculum, parietal operculum, postcentral gyrus, Heschl gyrus, superior temporal gyrus, parietal superior and inferior, supramarginal and angular gyri, cuneus, precuneus, paracentral lobule, superior and middle occipital gyri. Superior and posterior part corona radiata, superior longitudinal fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus, superior temporal gyrus. Caudate, putamen. Anterior limb of internal capsule, anterior, superior and posterior corona radiata, external capsule, superior longitudinal fasciculus, superior fronto-occipital fasciculus. Frontal operculum, middle and inferior frontal gyri, insula, Heschl gyrus, superior and middle temporal gyri, parietal operculum. Putamen. Anterior limb of internal capsule, external capsule. Middle and inferior frontal gyri. Anterior part of corona radiata, superior fronto-occipital fasciculus Insula, parietal operculum, Heschl gyrus. Hippocampus, putamen, pallidum, thalamus. Cerebral peduncle, posterior limb internal capsule, rentrolenticular part of internal capsule, superior corona radiata, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus. Caudate, putamen, pallidum. Anterior and posterior limb of internal capsule, anterior and superior part of corona radiata, external capsule, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus. Precentral gyrus, frontal operculum, insula, parietal operculum, postcentral gyrus, Heschl gyrus, superior temporal gyrus, supramarginal gyrus. Caudate, putamen, pallidum, thalamus. Anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior, superior and posterior corona radiata, sagittal stratum, external capsule, superior longitudinal fasciculus, superior frontooccipital fasciculus, uncinate fasciculus. Precentral gyrus, frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus, superior temporal gyrus. Hippocampus, amydgala, caudate, putamen, pallidum, thalamus. Cerebral peduncle, anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior, superior and posterior corona radiata, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus. Putamen, pallidum. Anterior limb of internal capsule, anterior part of corona radiata, external capsule, superior fronto-occipital fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, postcentral gyrus, posterior cingulum, Heschl gyrus, superior, middle and inferior temporal gyrus, fusiform gyrus, parietal inferior, supramarginal and angular gyri, splenium of corpus callosum, precuneus, cuneus, lingual gyrus, calcarine fissure, superior, middle and inferior occipital gyri. Hippocampus, parahippocampal gyrus, putamen, pallidum, thalamus. Posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior and posterior part of corona radiata, posterior thalamic radiation, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus. Precentral gyrus, frontal operculum, inferior frontal gyrus, olfactory cortex, insula, parietal operculum, postcentral gyrus, fusiform gyrus, Heschl gyrus, superior, middle and inferior temporal gyri, parietal inferior, supramarginal and angular gyri. Hippocampus, parahippocampal gyrus, amygdala, caudate, putamen, pallidum, thalamus. Anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior, superior and posterior corona radiata, posterior thalamic radiation, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus. Precentral gyrus, superior and middle frontal gyri, supplementary motor area, postcentral gyrus, parietal superior and inferior, paracentral lobule. Superior part of corona radiata, superior longitudinal fasciculus Caudate, thalamus. Posterior limb of internal capsule, rentrolenticular part of internal capsule, superior and posterior corona radiata, sagittal stratum, external capsule, fornix, superior fronto-occipital fasciculus.

Brodmann ares involved

1, 2, 3, 4, 6, 22, 39, 40, 41, 42, 43

22, 39, 40, 41, 42 22, 38, 44, 45, 47

1, 2, 3, 7, 18, 19, 20, 21, 22, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 47

1, 2, 3, 4, 5, 6, 7, 39, 40, 41, 42, 43, 44

38, 45, 47

21, 22, 38, 44, 45, 47

9, 44, 46 20, 41

44, 45, 47

3, 6, 34, 41

34, 44

– 1, 2, 3, 17, 18, 19, 20, 21, 22, 27, 29, 30, 37, 40, 41, 42, 43, 44, 45, 47

1, 2, 3, 4, 20, 21, 22,28, 34, 36, 38, 39, 40, 41, 42, 43, 44, 45, 47

2, 3, 4, 6, 40, 43

–



12

Table 6 (continued ) Case MB

MR

PM

Regions affected by the lesion

Brodmann ares involved

Frontal operculum, middle and inferior frontal gyrus, insula, parietal operculum, postcentral gyrus, Heschl gyrus, superior and middle temporal gyri, supramarginal gyrus. Hippocampus, amygdala, caudate, putamen, pallidum. Anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior, superior and posterior corona radiata, posterior thalamic radiation, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus, superior temporal gyrus. Hippocampus, amygdala, putamen, pallidum, thalamus. Anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior and superior corona radiata, sagittal stratum, external capsule, fornix, superior longitudinal fasciculus, uncinate fasciculus. Frontal operculum, inferior frontal gyrus, insula, parietal operculum, Heschl gyrus, superior temporal gyrus. Amygdala, caudate, putamen, pallidum, thalamus. Anterior and posterior limb of internal capsule, rentrolenticular part of internal capsule, anterior, superior and posterior corona radiata, external capsule, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus.

21, 22, 34, 37, 38, 41, 42, 44, 45, 46, 47

34, 45, 47

34, 38, 45, 47

Fig. 6. Results of the lesion overlap for each group of patients. The overlap of LBD and RBD patients' lesions is shown. Lesions are displayed in radiological convention, the left side of the template shows the right hemisphere and vice versa.

they performed the spatial ideomotor subtask, as suggested by both lower accuracy and slower RTs. Moreover, RBD patients did not show the effect of spatial compatibility in the anatomical ideomotor subtask. Indeed, the effect of spatial compatibility was observed by both LBD patients and healthy participants who were slower in the non-compatible condition of the anatomical subtask. This can be due to the absence of the spatial matching between the participant and the model, or to a facilitatory effect in the compatible condition when the produced movement matches both spatially and anatomically that of the model. Studies on children showed that movements presented in a mirror perspective are easier to imitate (Wapner and Cirillo, 1968; Schofield, 1976; Bekkering et al., 2000; Gleissner et al., 2000). Indeed, in many circumstances anatomical imitation is more difficult than spatial imitation, because it might require the inhibition of the automatic tendency to mirror the model's movements and an additional spatial transformation of the perceived movements from the model's body to the imitator's body. Patients with frontal lesions find more difficult to imitate gestures in an anatomical than in a spatial perspective (Chiavarino et al., 2007), and the advantage of spatial imitation has been shown also in healthy adults (Avikainen et al., 2003). Moreover, in previous studies in which a paradigm similar to the ideomotor task used in the present study was employed, healthy participants were faster at imitating in the spatial than in anatomical condition (Mengotti et al., 2012, 2013b). Therefore, the difference in performance between LBD and RBD patients (i.e., the anatomical subtask being more difficult than the spatial subtask for the LBD but not for the

Fig. 7. Results of the subtraction analysis for the ideomotor task. Regions specifically damaged in the group of patients that presented a deficit in the condition of interest are shown, after subtraction from the group of patients that did not present a deficit in that condition. In (B) and (C) subtraction maps for compatible and noncompatible conditions were then superimposed only for display purposes, in these maps the compatible condition is shown in red, the non-compatible condition in blue and the overlap between the two conditions is shown in violet. Lesions are displayed in radiological convention, the left side of the template shows the right hemisphere and vice versa. STG¼ superior temporal gyrus; OP¼ parietal operculum; SLF ¼ superior longitudinal fasciculus; IFG ¼inferior frontal gyrus; FX ¼fornix; SCR ¼ corona radiata superior part. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

RBD) argues in favor of a differential processing of information in the two hemispheres in imitation. In addition, spatial imitation is not always easier and more automatic than anatomical imitation. For instance, when the model is asked to move always the same hand, or to perform a



Fig. 8. Results of the subtraction analysis for the gesture imitation task. Regions specifically damaged in the group of patients that presented a deficit in the condition of interest are shown, after subtraction from the group of patients that did not present a deficit in that condition. Subtraction maps for anatomical and spatial conditions were then superimposed only for display purposes, the anatomical condition is shown in red, the spatial condition in blue and the overlap between the two conditions is shown in violet. Lesions are displayed in radiological convention, the left side of the template shows the right hemisphere and vice versa. STG¼ superior temporal gyrus; SCR ¼corona radiata superior part; IFG¼ inferior frontal gyrus; SMG ¼ supramarginal gyrus; ANG ¼ angular gyrus; ECP¼ external capsule; ICP ¼ internal capsule; PMC ¼premotor cortex; SLF ¼superior longitudinal fasciculus; M1 ¼primary motor cortex. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

movement always towards the same target, spatial imitation is preferred in the former condition and anatomical imitation in the latter condition (Franz et al., 2007). Moreover, Sartori et al. (2013) showed that when participants observed movements performed by a model with either the left or right hand, the effect of motor resonance, i.e. higher motor-evoked potentials in the muscles of the hands of participants when they are observing hand movements, was not present in the hand spatially compatible with the model (i.e. favoring the mirror perspective). Such effect was observed in the dominant hand, irrespective of the fact that this preactivation of the motor system was matching anatomically or spatially the movement of the model. Taken together these results suggest that while the anatomical and spatial imitation may both be necessary for a refined imitative performance, the spatial imitation and the anatomical imitation are more useful, respectively, when the focus of attention is on the final target of the movement, and when a more accurate processing of the effector or of the kinematics of the movement is needed (Franz et al., 2007). These different mechanisms seem to be mapped differently in the two hemispheres, with the left hemisphere being more involved in processing information related to the body of the model, and the right hemisphere mainly coding for the spatial position of the movements. The investigation of anatomical and spatial components of imitation is critical also in clarifying the role of the right hemisphere in imitation. Apraxic patients with right hemisphere damage have been reported (De Renzi et al., 1980; Goldenberg, 1999, 2009; Della Sala et al., 2006; Tessari et al., 2007), even if the frequency of the disorder is lower than following left hemisphere damage. In particular, RBD patients show higher difficulties than LBD patients in imitating meaningless gestures (Tessari et al., 2007) and finger configurations (Goldenberg, 1999, 2009; Della Sala et al., 2006). Our results confirm previous findings. Indeed, RBD patients were more impaired in imitating meaningless

13

gestures in both anatomical and spatial conditions. Moreover the error analysis showed that RBD patients were more prone to errors with finger postures than LBD patients, thus their performance in the ideomotor task, in which they are asked to reproduce finger movements, can partially reflect this selective deficit. RBD patients' more impaired performance of meaningless gestures and finger postures, together with a general impairment at coding the spatial component in both ideomotor and gesture imitation tasks, support the idea that the right hemisphere is competent for visuospatial abilities including the visuo-spatial analysis of gestures (Goldenberg, 1999). In the present study, RBD patients were impaired in both anatomical and spatial conditions in the gesture imitation task, however their reduced performance in the anatomical condition can be explained by the presence of neglect in some patients. In fact, in this condition the gestures were shown in the left side of the space. However, the difficulties in gesture processing observed in RBD patients throughout the present and the past studies (Della Sala et al., 2006; Goldenberg, 2009) cannot be simply explained by the presence of neglect. Patients with neglect tend to commit a higher proportion of errors than those without neglect, in case of severe neglect, nevertheless also patients without neglect were found to be impaired in imitating finger postures (Della Sala et al., 2006; Goldenberg, 2009) or meaningless gestures (Tessari et al., 2007). LBD patients experienced difficulties in the anatomical imitation, but only in the conditions in which they have to rely solely on the anatomical matching with the model, whereas they performed normally on the anatomical imitation when model's and patients' movements matched also at the spatial level. This is shown in the ideomotor task: LBD patients were impaired only in the noncompatible condition of the anatomical subtask, that is the only condition in which the spatial matching between model and imitator is not present, suggesting an inability to perform anatomical imitation in the absence of any other similarity, e.g., the spatial one, between their movements and the model's movements. They also showed impairment in performing gesture imitation only in the anatomical condition, thus suggesting a facilitatory effect of the mirror (i.e., spatial) perspective and a difficulty in performing the additional spatial transformation of the perceived movements required by this type of imitation. As for the gesture imitation task, our results do not completely fulfill our predictions about a dissociating pattern between LBD (with impaired performance in the anatomical condition but not in the spatial condition) and RBD patients (with impaired performance in the spatial condition but not in the anatomical condition). Indeed, RBD patients showed impaired performance in both conditions; however we believe that the presence of neglect in some of the patients might have been detrimental for their performance in the anatomical condition. As shown in our descriptive analysis, when only non-neglect patients are considered, the improvement of the performance in the anatomical condition seems to be more consistent than the improvement in the spatial condition. Moreover, we cannot completely rule out the possibility that the performance of LBD patients in the anatomical condition might be partially affected by an additional spatial transformation required to complete this condition, in order to match the imitator's with the model's body. However, some previous results suggest that anatomical imitation does not automatically involve this additional spatial translation towards the model's body (Franz et al., 2007), therefore leaving the issue open. The results of the present study provide useful information about the way in which the matching between perception and action is achieved. In normal conditions, the anatomical identity of the body parts moved by the model and the spatial locations of the movements in space are integrated and used to perform imitation


14


(Mengotti et al., 2012). Following brain damage, sometimes only the anatomical or the spatial imitation is possible. The lesion analysis allowed us to identify the neural underpinnings of the anatomical and spatial components of imitation. Considering the ideomotor task, LBD patients showed differences in the regions associated with impaired performance in either the anatomical or the spatial subtask. The lesion to the parietal operculum is primarily responsible for the impaired performance in the anatomical subtask (non-compatible condition). This result is consistent with previous evidence that associated the activation of the parietal operculum with the processing of the anatomical component of imitation (Mengotti et al., 2012). However the opercular area highlighted in the subtraction analysis is slightly different from that revealed by the same task in fMRI (Mengotti et al., 2012); the fMRI study indeed suggested an involvement of the cytoarchitectonical area OP4, whereas in the present study areas OP1 and OP2 were found to be involved. Both OP4 and OP1 are part of the secondary somatosensory region and are known to respond to tactile stimulation with somatotopic organization (Eickhoff et al., 2007). In particular, the hand representation of the somatotopic map extends in both areas, suggesting that the somatosensory representation can be used in imitation, when actions that involve hands are presented. However, OP1 and OP4 have been shown to differ in connectivity, with OP4 being more connected with motor and premotor cortices, and OP1 being more connected with the anterior parietal cortex and thalamus (Eickhoff et al., 2010). Interestingly, OP1 was found to be associated with the processing of the body representations (Corradi-Dell’Acqua et al., 2009), suggesting that a connection with body representations is present when imitating anatomically (i.e., in the non-compatible condition). For the spatial subtask, lesion analysis revealed the involvement of the BA 44, a part of the Broca's area, repeatedly associated with imitation (Heiser et al., 2003; Watanabe et al., 2011; but see Makuuchi, 2005 for a different result). Our results seem to suggest a role of the BA 44 in imitation when the model is presented in a mirror perspective and therefore a direct matching between model and imitator was present. Nevertheless, since only a few LBD patients were impaired in the spatial subtask, this result has to be considered only as suggestive of an involvement of this region in spatial imitation, and more research is needed to confirm this result. The analysis of RBD patients' lesions revealed the involvement of the parietal operculum in all conditions of the ideomotor task. Differently from LBD patients, the OP3 in RBD patients seems to be extensively recruited in all conditions. OP3 is the more ventral part of the parietal operculum, and is considered the human homolog of the “ventral somatosensory area” in non-human primates. Similarly to OP1 and OP4, OP3 contains a somatotopic map sensitive to tactile stimuli (Eickhoff et al., 2007). It is worth noting that, in our study, lesions to this area led to defective performance in the ideomotor task in RBD patients irrespective of the matching required with the model, whether based on the anatomical identity of the body parts moved or on the spatial location in space of the movements. The involvement of the right parietal operculum is in line with previous evidence showing a bilateral involvement of this region in imitation (Mengotti et al., 2012, 2013b). The drop in gesture imitation of LBD patients seems to be caused by parietal lesions, involving the supramarginal and the angular gyri, and part of the temporal cortex. Left parietal regions are traditionally associated with the presence of deficits in praxis (Goldenberg and Karnath, 2006; Tessari et al., 2007; Mengotti et al., 2013a), thus our results are in line with the previous literature. The regions associated with the anatomical and the spatial component seem to be the same, but the spatial component involved a more extended part of the parietal cortical surface than the anatomical component, suggesting that only larger lesions can

impair the spatial imitation. The distribution of regions that are associated with more impaired imitative performance in RBD patients is different with respect to that of LBD patients, and primarily involves the premotor region, suggesting an important difference in the recruitment of brain regions involved in imitation in the left and right hemisphere. Also when imitating gestures, RBD patients seem to rely on somatosensory representations, as lesions to the primary somatosensory cortex led to defective performance in the gesture imitation task, but only in the anatomical condition. A comparison between the two tasks used in the present study leads to the conclusion that both tasks are useful to investigate in the ability to perform imitation on the basis of the anatomical or on the spatial matching with the model. The general impairment in the spatial component of imitation for RBD patients is shown in both tasks, as well as the increased difficulty of LBD patients with the processing of the anatomical component. Nevertheless, the brain regions involved in the two tasks are different. The ideomotor task recruited mainly frontal regions and the parietal opercula, in line with past neuroimaging studies that used similar tasks (Iacoboni et al., 1999; Brass et al., 2001b; Koski et al., 2003; Bien et al., 2009; Mengotti et al., 2012), whereas the gesture imitation task involved parietal regions, confirming previous neuropsychological evidence (Haaland et al., 2000; Weiss et al., 2001; Tessari et al., 2007; Mengotti et al. 2013a). Therefore, our results point to an anatomo-functional distinction between the processes activated in such tasks. Indeed, even if both tasks are informative for studying the cognitive processes that support the different types of matching in imitation, the underlying differences of the two tasks have to be taken into account when testing patients' performance. The results of the present study speak to the correspondence problem. Indeed, our data show how the matching between model's and imitator's movements can be achieved on the basis of either the anatomical or the spatial similarity between the movements, with hemispheric specializations. Even if the action of mirroring the model's movements seems to be easy and automatic, also information concerning the anatomical matching with the model is relevant for efficiently replicating the movements. We now address some methodological limitations of the current study. First, our sample did not include many apraxic patients. Using the cut-off of the imitation test that we employed in the present paper to diagnose the presence of ideomotor apraxia (Tessari et al., 2015), two out of 10 LBD patients and two out of 10 RBD patients would be considered as apraxics. However when the patients' performance is compared with the group of controls matched for age and education (as showed in Table 4 in which single patients' t scores are presented), six LBD patients showed lower performance than controls in at least one of the two gesture imitation conditions, and seven RBD patients showed lower performance than controls in at least one of the two gesture imitation conditions. Even if patients' performance did not fall below the cut-off on the standardized test (Tessari et al., 2015), they however showed mild deficits in performing gesture imitation. Moreover, as in the standardized test gestures only tap the spatial component, apraxic patients would show, by definition, impaired spatial matching. If only the standard test for apraxia were applied, we would not be able to identify dissociations between the anatomical and spatial matching components. In fact, apraxic patients can show either a selective deficit in the spatial matching or a deficit in both the anatomical and spatial components. Moreover, for the same reason, the performance of non-apraxic patients is of particular importance to the aim of our study, as these patients might reveal a selective deficit of the anatomical component, or subtle deficits in both anatomical and spatial matching that contribute to a less effective imitative performance without leading



necessarily to ideomotor apraxia. Our findings are not specific for apraxia and provide further insight about the contribution of the anatomical and spatial matching processes to the imitative performance. Second, in the gesture imitation task, we cannot completely exclude that the results found could be, at least partially, attributable to the presence of two potentially confounding factors. One is the presence of neglect in some of the RBD patients, and another is the putative additional spatial transformation required to perform the anatomical condition that concerns, in particular, the LBD patients' performance. With respect to the former confounding factor, we found that the presence of neglect was correlated with the scores in the anatomical condition of the gesture imitation task: the more severe the neglect, the lower the scores in the anatomical condition. Therefore, the RBD group did not show the expected dissociation between anatomical and spatial gesture imitation. However, if we consider the RBD patients without neglect, their group performance improves in the anatomical condition, in line with our predictions. As for the latter confounding factor, our results cannot clarify whether these transformations are actually required to accomplish the task in the anatomical condition. Future studies are needed to elucidate this issue.

5. Conclusions By analyzing the performance of left and right brain-damaged patients in two different imitation tasks, the present study brings evidence in support of the ideomotor view of imitation, suggesting that imitation is achieved through shared representations between perception and action production. Therefore imitation is based on the similarity between percepts and performed movements, and this similarity can be based either on the anatomical or on the spatial matching between model and imitator. On the one hand, lesions in the left hemisphere are found to impair the processing of the anatomical component, especially when it has to be processed without the concurrent presence of the spatial matching between model's and imitator's movements. On the other hand, lesions in the right hemisphere lead to a general impairment in processing the spatial component of imitation, supporting the view of an additional contribution of the right hemisphere in processing the visuo-spatial features of the gestures.

Acknowledgments We would like to thank Dr. Alberta Lunardelli, Dr. Alina Menichelli and Dr. Francesca Meneghello for their help with the neuropsychological evaluations and Dr. Marilena Aiello and Dr. Elisabetta Ambron for their precious help with the analysis of the patients' recordings. Moreover, we would like to thank also Irene Cogliati Dezza for her help in testing some of the control participants. This research was supported by a Grant (PRIN) awarded to R.I.R. by the Italian Ministry of Education, University and Research and by a post-doctoral fellowship awarded to P.M. by the FSE (Fondo Sociale Europeo) of the Friuli-Venezia Giulia Government (S.H.A.R. M. project-Supporting Human Assets in Research and Mobility) in collaboration with S.I.D.EM. S.p.A.

References Avikainen, S., Wohlschläger, A., Liuhanen, S., Hänninen, R., Hari, R., 2003. Impaired mirror-image imitation in Asperger and high-functioning autistic subjects. Curr. Biol. 13 (4), 339–341.

15

Bekkering, H., Wohlschläger, A., Gattis, M., 2000. Imitation of gestures in children is goal-directed. Q. J. Exp. Psychol.: Sect. A 53 (1), 153–164. Bertenthal, B.I., Longo, M.R., Kosobud, A., 2006. Imitative response tendencies following observation of intransitive actions. J. Exp. Psychol.: Hum. Percept. Perform. 32 (2), 210–225. Bien, N., Roebroeck, A., Goebel, R., Sack, A.T., 2009. The brain's intention to imitate: the neurobiology of intentional versus automatic imitation. Cereb. Cortex 19 (10), 2338–2351. Boyer, T.W., Longo, M.R., Bertenthal, B.I., 2012. Is automatic imitation a specialized form of stimulus–response compatibility? dissociating imitative and spatial compatibilities. Acta Psychol. 139 (3), 440–448. Brass, M., Bekkering, H., Wohlschläger, A., Prinz, W., 2000. Compatibility between observed and executed finger movements: comparing symbolic, spatial, and imitative cues. Brain Cogn. 44 (2), 124–143. Brass, M., Bekkering, H., Prinz, W., 2001a. Movement observation affects movement execution in a simple response task. Acta Psychol. 106 (1), 3–22. Brass, M., Zysset, S., von Cramon, D.Y., 2001b. The inhibition of imitative response tendencies. NeuroImage 14 (6), 1416–1423. Brass, M., Derrfuss, J., Matthes-von Cramon, G., Von Cramon, D.Y., 2003. Imitative response tendencies in patients with frontal brain lesions. Neuropsychology 17 (2), 265–271. Brass, M., Heyes, C., 2005. Imitation: is cognitive neuroscience solving the correspondence problem? Trends Cogn. Sci. 9 (10), 489–495. Caffarra, P., Vezzadini, G., Dieci, F., Zonato, F., Venneri, A., 2002. Rey–Osterrieth complex figure: normative values in an Italian population sample. Neurol. Sci. 22 (6), 443–447. Capasso, R., Miceli, G., 2001. Esame Neuropsicologico per Pazienti Afasici (ENPA). Springer-Verlag, Milano. Carlesimo, G.A., Caltagirone, C., Gainotti, G., 1996. The mental deterioration battery: normative data, diagnostic reliability and qualitative analyses of cognitive impairment. Eur. Neurol. 36 (6), 378–384. Carmo, J.C., Rumiati, R.I., 2009. Imitation of transitive and intransitive actions in healthy individuals. Brain Cogn. 69 (3), 460–464. Chiavarino, C., Apperly, I.A., Humphreys, G.W., 2007. Exploring the functional and anatomical bases of mirror-image and anatomical imitation: the role of the frontal lobes. Neuropsychologia 45 (4), 784–795. Corradi-Dell’Acqua, C., Tomasino, B., Fink, G.R., 2009. What is the position of an arm relative to the body? neural correlates of body schema and body structural description. J. Neurosci. 29 (13), 4162–4171. Crawford, J.R., Garthwaite, P.H., 2005. Testing for suspected impairments and dissociations in single-case studies in neuropsychology: evaluation of alternatives using Monte Carlo simulations and revised tests for dissociations. Neuropsychology 19 (3), 318–331. Cubelli, R., Marchetti, C., Boscolo, G., Della Sala, S., 2000. Cognition in action: testing a model of limb apraxia. Brain Cogn. 44 (2), 144–165. De Renzi, E., Faglioni, P., 1999. Apraxia. In: Denes, G., Pizzamiglio, L. (Eds.), Handbook of Clinical and Experimental Neuropsychology. Psychology Press, Hove, UK, pp. 421–440. De Renzi, E., Lucchelli, F., 1988. Ideational apraxia. Brain 111, 1173–1185. De Renzi, E., Motti, F., Nichelli, P., 1980. Imitating gestures: a quantitative approach to ideomotor apraxia. Arch. Neurol. 37 (1), 6–10. Della Sala, S., Faglioni, P., Motto, C., Spinnler, H., 2006. Hemisphere asymmetry for imitation of hand and finger movements, Goldenberg’s hypothesis reworked. Neuropsychologia 44 (8), 1496–1500. Eickhoff, S.B., Stephan, K.E., Mohlberg, H., Grefkes, C., Fink, G.R., Amunts, K., Zilles, K., 2005. A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. NeuroImage 25 (4), 1325–1335. Eickhoff, S.B., Grefkes, C., Zilles, K., Fink, G.R., 2007. The somatotopic organization of cytoarchitectonic areas on the human parietal operculum. Cereb. Cortex 17 (8), 1800–1811. Eickhoff, S.B., Jbabdi, S., Caspers, S., Laird, A.R., Fox, P.T., Zilles, K., Behrens, T.E., 2010. Anatomical and functional connectivity of cytoarchitectonic areas within the human parietal operculum. J. Neurosci. 30 (18), 6409–6421. Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12 (3), 189–198. Franz, E.A., Ford, S., Werner, S., 2007. Brain and cognitive processes of imitation in bimanual situations: making inferences about mirror neuron systems. Brain Res. 1145, 138–149. Giovagnoli, A.R., Del Pesce, M., Mascheroni, S., Simoncelli, M., Laiacona, M., Capitani, E., 1996. Trail making test: normative values from 287 normal adult controls. Ital. J. Neurol. Sci. 17, 305–309. Gleissner, B., Bekkering, H., Meltzoff, A.N., 2000. Children's coding of human action: cognitive factors influencing imitaation in 3‐year‐old. Dev. Sci. 3 (4), 405–414. Goldenberg, G., 1995. Imitating gestures and manipulating a mannikin—the representation of the human body in ideomotor apraxia. Neuropsychologia 33 (1), 63–72. Goldenberg, G., 1999. Matching and imitation of hand and finger posturesin patients with damage in the left or right hemispheres. Neuropsychologia 37 (5), 559–566. Goldenberg, G., Karnath, H.O., 2006. The neural basis of imitation is body part specific. J. Neurosci. 26 (23), 6282–6287. Goldenberg, G., 2009. Apraxia and the parietal lobes. Neuropsychologia 47 (6), 1449–1459. Goldenberg, G., Münsinger, U., Karnath, H.O., 2009. Severity of neglect predicts accuracy of imitation in patients with right hemisphere lesions.


16


Neuropsychologia 47 (13), 2948–2952. Haaland, K.Y., Harrington, D.L., Knight, R.T., 2000. Neural representations of skilled movement. Brain 123 (11), 2306–2313. Heiser, M., Iacoboni, M., Maeda, F., Marcus, J., Mazziotta, J.C., 2003. The essential role of Broca's area in imitation. Eur. J. Neurosci. 17 (5), 1123–1128. Heyes, C., Ray, E., 2004. Spatial SR compatibility effects in an intentional imitation task. Psychon. Bull. Rev. 11 (4), 703–708. Holmes, C.J., Hoge, R., Collins, L., Woods, R., Toga, A.W., Evans, A.C., 1998. Enhancement of MR images using registration for signal averaging. J. Comput. Assist. Tomogr. 22, 324–333. Hommel, B., Müsseler, J., Aschersleben, G., Prinz, W., 2001. The theory of event coding (TEC): a framework for perception and action planning. Behav. Brain Sci. 4, 849–937. Iacoboni, M., Woods, R.P., Brass, M., Bekkering, H., Mazziotta, J.C., Rizzolatti, G., 1999. Cortical mechanisms of human imitation. Science 286 (5449), 2526–2528. Koski, L., Iacoboni, M., Dubeau, M.C., Woods, R.P., Mazziotta, J.C., 2003. Modulation of cortical activity during different imitative behaviors. J. Neurophysiol. 89 (1), 460–471. Liepmann, H., 1920. Apraxie [Apraxia]. Ergeb. Gesamten Med. 1, 516–543. Longo, M.R., Bertenthal, B.I., 2009. Attention modulates the specificity of automatic imitation to human actors. Exp. Brain Res. 192 (4), 739–744. Longo, M.R., Kosobud, A., Bertenthal, B.I., 2008. Automatic imitation of biomechanically possible and impossible actions: effects of priming movements versus goals. J. Exp. Psychol.: Hum. Percept. Perform. 34 (2), 489–501. Luzzatti, C., Willmes, K., De Bleser, R., 1996. Aachener Aphasie Test: Versione Italiana [Aachener Aphasia Test: Italian version]. Organizzazioni Speciali, Firenze, Italy. Makuuchi, M., 2005. Is Broca's area crucial for imitation? Cereb. Cortex 15 (5), 563–570. Massen, C., Prinz, W., 2009. Movements, actions and tool-use actions: an ideomotor approach to imitation. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 364, 2349–2358. Mauri, M., Carlesino, G.A., Graceffa, A.M.S., Loasses, A., Lorusso, S., Sinforiani, E., Bono, G., Caltagirone, C., 1997. Standardizzazione di due nuovi test di memoria: apprendimento di liste di parole correlate e non correlate semanticamente. Arch. Psicol. Neurol. Psichiatr. 58, 621–645. Mengotti, P., Corradi-Dell’Acqua, C., Rumiati, R.I., 2012. Imitation components in the human brain: an fMRI study. NeuroImage 59 (2), 1622–1630. Mengotti, P., Corradi-Dell’Acqua, C., Negri, G.A., Ukmar, M., Pesavento, V., Rumiati, R.I., 2013a. Selective imitation impairments differentially interact with language processing. Brain 136 (8), 2602–2618. Mengotti, P., Ticini, L.F., Waszak, F., Schütz‐Bosbach, S., Rumiati, R.I., 2013b. Imitating others' actions: transcranial magnetic stimulation of the parietal opercula reveals the processes underlying automatic imitation. Eur. J. Neurosci. 37 (2), 316–322. Menz, M.M., McNamara, A., Klemen, J., Binkofski, F., 2009. Dissociating networks of imitation. Hum. Brain Mapp. 30 (10), 3339–3350. Oldfield, R.C., 1971. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9, 97–113. Orsini, A., Grossi, D., Capitani, E., Laiacona, M., Papagno, C., Vallar, G., 1987. Verbal and spatial immediate memory span: normative data from 1355 adults and

1112 children. Ital. J. Neurol. Sci. 8 (6), 537–548. Papagno, C., Della Sala, S., Basso, A., 1993. Ideomotor apraxia without aphasia and aphasia without apraxia: the anatomical support for a double dissociation. J. Neurol. Neurosurg. Psychiatry 56 (3), 286–289. Prinz, W., 1997. Perception and action planning. Eur. J. Cogn. Psychol. 9 (2), 129–154. Riddoch, M.J., Humphreys, G.W., 1993. BORB: Birmingham Object Recognition Battery. LEA, Hove, UK. Rorden, C., Brett, M., 2000. Stereotaxic display of brain lesions. Behav. Neurol. 12 (4), 191–200. Rorden, C., Karnath, H.O., 2004. Using human brain lesions to infer function: a relic from a past era in the fMRI age? Nat. Rev. Neurosci. 5 (10), 812–819. Rothi, L.J.G., Ochipa, C., Heilman, K.M., 1991. A cognitive neuropsychological model of limb praxis. Cogn. Neuropsychol. 8 (6), 443–458. Rumiati, R.I., Weiss, P.H., Tessari, A., Assmus, A., Zilles, K., Herzog, H., Fink, G.R., 2005. Common and differential neural mechanisms supporting imitation of meaningful and meaningless actions. J. Cogn. Neurosci. 17 (9), 1420–1431. Rumiati, R.I., Papeo, L., Corradi‐Dell’Acqua, C., 2010. Higher‐level motor processes. Ann. NY Acad. Sci. 1191 (1), 219–241. Sartori, L., Begliomini, C., Castiello, U., 2013. Motor resonance in left-and righthanders: evidence for effector-independent motor representations. Front. Hum. Neurosci. 7, 33. Schofield, W.N., 1976. Hand movements which cross the body midline: findings relating age differences to handedness. Percept. Motor Skills 42, 643–646. Spinnler, H., Tognoni, G., 1987. Standardizzazione e taratura italiana di test neuropsicologici [Italian standardization and validation neuropsychological tests]. Ital. J. Neurol. Sci. 6, 1–20. Tessari, A., Rumiati, R.I., 2004. The strategic control of multiple routes in imitation of actions. J. Exp. Psychol.: Hum. Percept. Perform. 30 (6), 1107–1116. Tessari, A., Canessa, N., Ukmar, M., Rumiati, R.I., 2007. Neuropsychological evidence for a strategic control of multiple routes in imitation. Brain 130 (4), 1111–1126. Tessari, A., Toraldo, A., Lunardelli, A., Zadini, A., Rumiati, R.I., 2015. STIMA: a short screening test for ideo-motor apraxia, selective for action meaning and bodily district. Neurol. Sci., 1–8. Wapner, S., Cirillo, L., 1968. Imitation of a model’s hand movements: age changes in transposition of left-right relations. Child Dev. 39 (3), 887–894. Warrington, E.K., James, M., 1991. The Visual Object and Space Perception Battery. Thames Valley Test Company, Bury St Edmunds, UK. Warrington, E.K., 1996. The Camden Memory Test. Psychology Press, Hove, UK. Watanabe, R., Watanabe, S., Kuruma, H., Murakami, Y., Seno, A., Matsuda, T., 2011. Neural activation during imitation of movements presented from four different perspectives: a functional magnetic resonance imaging study. Neurosci. Lett. 503 (2), 100–104. Weiss, P.H., Dohle, C., Binkofski, F., Schnitzler, A., Freund, H.J., Hefter, H., 2001. Motor impairment in patients with parietal lesions: disturbances of meaningless arm movement sequences. Neuropsychologia 39 (4), 397–405. Wilson, B., Cockburn, J., Halligan, P., 1987. Behavioural Inattention Test. Thames Valley Test Company, Fareham, Hants, England.


Anatomical left-right asymmetry of language-related temporal cortex is different in left- and right-handers.

Effects of stimulus degradation on letter-matching performance of left and right hemispheric stroke patients.

Spatial ability in subgroups of left- and right-handers.

The right time and the left time: Spatial associations of temporal cues affect target detection in right brain-damaged patients.

Left and right lateralization for letter matching: strategy and sex differences.

A study of right unilateral spatial neglect in left hemispheric lesions: the difference between right-handed and non-right-handed post-stroke patients.

Multivariate pattern analysis reveals anatomical connectivity differences between the left and right mesial temporal lobe epilepsy.

Effect of Left, Right and Alternate Nostril Breathing on Verbal and Spatial Memory.

Effects of left-right orientation and position reversals on spatial perspective taking in young children.

Telling left from right: left-right asymmetric controls in sea urchins.

COMPARATIVE ANATOMICAL STUDY BETWEEN THE RIGHT AND LEFT SIDES OF THE AXILLARY NERVE IN RELATION TO DELTOPECTORAL APPROACH AND ACROMION.

Left-right spatial agnosia and other mental defects that characterize clinical researchers.

Differing emotional response from right and left hemispheres.

Multiple feeding vessels from left circumflex artery and right coronary artery to myxoma in left atrium.

Matching the right dog to the right home.

Super-dominant right coronary artery, absent left circumflex artery and left anterior descending artery arising from right coronary sinus.

The influence of stimulus proximity on judgments of spatial relationships in patients with chronic unilateral right or left hemisphere stroke.

Anatomical Closure of Left-to-Right Shunts in Premature Infants with Bronchopulmonary Dysplasia and Pulmonary Hypertension: A Cautionary Tale.

Anatomical Study of Chiari Network and the Remnant of Left Venous Valve in the Interior of Right Atrium.

Genetic factors in coronary heart-disease: Anatomical evidence from mice.

Aphasia in left-handers. Comparison of aphasia profiles and language recovery in non-right-handed and matched right-handed patients.

Dyscalculia, dysgraphia, and left-right confusion from a left posterior peri-insular infarct.

Double left anterior descending coronary artery originating from left main coronary stem and right coronary artery.

Anomalous Left Coronary Artery Arising from Right Pulmonary Artery and Giant Left Atrium.