Elsevier

Brain and Cognition

Volume 53, Issue 3, December 2003, Pages 483-494
Brain and Cognition

Pantomimes are special gestures which rely on working memory

https://doi.org/10.1016/S0278-2626(03)00209-4Get rights and content

Abstract

The case of a patient is reported who presented consistently with overt deficits in producing pantomimes in the absence of any other deficits in producing meaningful gestures. This pattern of spared and impaired abilities is difficult to reconcile with the current layout of cognitive models for praxis. This patient also showed clear impairment in a dual-task paradigm, a test taxing the co-ordination aspect of working memory, though performed normally in a series of other neuropsychological measures assessing language, visuo-spatial functions, reasoning function, and executive function. A specific working memory impairment associated with a deficit of pantomiming in the absence of any other disorders in the production of meaningful gestures suggested a way to modify the model to account for the data. Pantomimes are a particular category of gestures, meaningful, yet novel. We posit that by their very nature they call for the intervention of a mechanism to integrate and synthesise perceptual inputs together with information made available from the action semantics (knowledge about objects and functions) and the output lexicon (stored procedural programmes). This processing stage conceived as a temporary workspace where gesture information is actively manipulated, would generate new motor programmes to carry out pantomimes. The model of gesture production is refined to include this workspace.

Introduction

The purposive movements hampered in apraxia include different types of gestures: transitive, involving tool use; intransitive (or symbolic); pantomimes, the mime of tool use; and meaningless (or novel). Several apraxia taxonomies have been proposed (see for a review De Renzi & Faglioni, 1999) based on the categories of gestures.

However, pantomimes presented a problem to this approach. For example, Morlaas (1928) distinguished between a deficit of transitive gestures, which he assimilated to ideational apraxia, and a deficit of intransitive gestures, matched to ideomotor apraxia; yet this dichotomy is ambiguous for pantomimes. They could be classified as transitive gestures, since objects are somehow involved, but could also be thought of as intransitive, since objects are not present during the miming. Indeed, some contemporary authors have suggested the use of pantomimes to assess ideomotor apraxia (e.g., Heilman, Rothi, & Valenstein, 1982), while others have used them to examine ideational apraxia (e.g., De Renzi, Faglioni, & Sorgato, 1982). Pantomimes have been listed among the diagnostic tools for ideomotor apraxia (e.g., Andrewes, 2001), yet correlations have been reported between pantomimes and tool use (e.g., Foundas et al., 1995; Goldenberg & Hagmann, 1997; Hécaen, 1978).

The classic dichotomy ideomotor/ideational apraxia is an oversimplification (for a discussion, see Buxbaum, 2001; Goldenberg, 2003). To circumvent it, recently cognitive models of praxis processing have been proposed (Cubelli, Marchetti, Boscolo, & Della Sala, 2000; Rothi et al., 1991, Rothi et al., 1997) in analogy with the models devised for language (Coltheart, Curtis, Atkins, & Haller, 1993; Patterson & Shewell, 1987). These models depict two independent processing routes, one dealing with the selection of meaningful gestures to be retrieved from long term memory, the other responsible for the on-line assembly and the execution of novel gestures. Within the context of studies on apraxia, pantomimes of objects use should not be confused with symbolic gestures describing objects use. The two gestures may be very similar. However, in some instances they differ, as in the case of scissors, the symbolic representation of which would require the rhythmic overlaying of the extended index on the extended middle finger, whilst their correct pantomime would imply a paced adjoining of the flexed thumb and index. Hence, pantomimes are meaningful gestures rarely performed in everyday life, which are therefore novel. Therefore, no motor programmes would be readily available from the long-term stores to produce pantomimes.

In this paper, we tackle the issue of whether or not pantomimes are a special category of gestures. If they were, could selective deficit of pantomimes be observed? How could the cognitive models for praxis account for such deficit?

The models distinguish between a lexical route responsible for the production and the imitation of meaningful (familiar) gestures, and a non-lexical route, assumed to be responsible for the imitation of all seen gestures, familiar and non-familiar alike (see Fig. 1).

The “form” and the content of the familiar gestures are thought to be stored in two long-term memory systems, the gestural lexicon and the action semantics. The lexicon is further divided into input and output. The input gestural lexicon stores the representations of all known gestures allowing for the recognition of familiar gestures. The output gestural lexicon contains the procedural knowledge for the production of known gestures. The semantic system (see Roy & Square, 1985) stores the knowledge about objects and tools, their function and the way in which they are used. The semantic system also stores the meaning of symbolic, intransitive (i.e., performed without the object) gestures, either iconic, which represent the shape of an object (e.g., binocular), or arbitrary (e.g., the military salute). The two systems deal with two different types of information. The semantic system allows one to “know” how a given object should be used, the lexicon contains the specific instructions allowing one to actually use it, e.g., one “knows” how to play a violin (and even be able to perform gestures representing the conventional description of a violin player), yet have no idea on how to really play it.

The lexical-semantic route is used for the imitation of meaningful gestures as well as for their production either spontaneously or elicited on command. The non-lexical route comprises a visuo-motor conversion mechanism involved in transcoding visual information into motor programmes. Hence, meaningful gestures could be imitated through both the lexical and the non-lexical route, whilst meaningless gestures could only be imitated via the non-lexical route. Cubelli et al. (2000) further maintained that the lexical and the non-lexical routes converge into a gestural memory buffer, aimed at holding the motor programmes until the gestures are executed.

Therefore, given the redundancy for the imitation of meaningful gestures embedded in the model, a selective impairment of the imitation of meaningless gestures is predicted by a deficit of the non-lexical route (see, e.g., the cases reported by Goldenberg & Hagmann, 1997). However, one could never expect to observe a selective sparing of the ability to imitate meaningless gestures.

Bartolo, Cubelli, Della Sala, Drei, and Marchetti (2001) reported on the case of MF, who presented with a clear impairment in the production and imitation of all kinds of meaningful gestures under all modality of request coupled with an impeccable ability to imitate meaningless gestures. The current layout of the model would run into difficulty in accounting for MF’s pattern of spared and impaired abilities.

Since MF’s imitation of meaningless gestures was flawless, both the visuo-motor conversion mechanism and the gestural buffer ought to be spared. Her deficit in producing and imitating meaningful gestures had to be traced back to a deficit located along the lexical route. The input gestural lexicon and the action semantics were spared as indicated by her normal performance in a series of gesture discrimination and identification tasks. Her impairment with meaningful gestures could only be accounted for by a deficit of the output lexicon or in accessing it, yet it was not clear why the spared non-lexical route did not permit the imitation of meaningful gestures.

Margolin (1984) proposed two ways by means of which it would be possible to correctly copy written words. One is the lexical procedure, whereby one reads the word and reproduces it as if writing under dictation. The other is the pictorial mode, using which the word is copied point to point as if it were a meaningless pattern. However, if the word to be copied is recognised as familiar the lexical route is activated by default, preventing the use of the pictorial strategy to be implemented. Similarly, the use of the non-lexical route to imitate a shown gesture recognised as familiar would be barred by the automatic activation of the input lexicon, leading to the selection of the correspondent motor programme within the gestural output lexicon.

Gestures that carry no meaning are processed only by visuo-motor conversion mechanism; in contrast, meaningful gestures would be processed only by the lexical route that is responsible for their production both on imitation and on command. Consequently, the deficit of production and imitation of familiar gestures would be invariably yoked following damage to either the action semantics or the output lexicon (Bartolo et al., 2001). However, should the input lexicon also be damaged (agnosia for known gestures, also labelled pantomime agnosia—Rothi, Mack, & Heilman (1986)) the production on command would still be impaired, whilst imitation of familiar gestures would be preserved being processed by the spared non-lexical route. The same dissociation between spared imitation and impaired production of familiar gestures could derive from a defective access to the action semantics from non-gestural inputs in the context of an otherwise intact lexical route (e.g., verbal commands which fail to activate the stored semantic representations).

The production of meaningful gestures, both transitive, involving tool use (e.g., the use of a hammer), and intransitive (i.e., symbolic, e.g., hitch-hiking), draws upon long-term representations either conceptual or procedural, sitting in the action semantics system and the output lexicon respectively. Transitive and intransitive gestures are represented separately both at the level of the action semantics and the output lexicon. The meaning of intransitive gestures varies according to the different socio-cultural contexts, whilst knowledge about transitive gestures is culture-free being conditioned by the object’s features. At the procedural level, motor programmes for intransitive gestures are independent of any environmental context while those for transitive gestures are less specified and conform to the physical attributes (shape, size or weight) of the objects. Thus it is plausible that intransitive and transitive gestures be selectively affected at both levels. Indeed Cubelli et al. (2000) reported on two patients with selective output lexicon impairment, limited to transitive (case 8) or to intransitive gestures (case 19). Similarly Ochipa, Rothi, and Heilman (1989) described the case of a patient whose deficits were restricted to tool knowledge.

Hughlings Jackson (1893) discussed pantomimes in the context of other symbolic gestures, suggesting a relation between pantomimes and intransitive gestures. Pantomimes do not imply the actual use of objects, rather they “represent something and presuppose the attitude of abstraction” (Goldstein, 1948, p. 137; Goodglass & Kaplan, 1963, p. 706) defined pantomimes as “improvised movements which … describe the physical properties of an object … or some action imposed by the object.” Wang and Goodglass (1992, p. 719) maintained that pantomimes were qualitatively different and more difficult to execute than over-learned gestures.

Performance in pantomime studies may depend upon the instructions given (see Raymer, Maher, Foundas, Heilman, & Rothi, 1997). For instance, Dumont, Ska, and Schiavetto (1999) asked their patients: “Show me how you would brush your teeth” (p. 450). Under these conditions participants are not asked explicitly to mime the use of a real object and they may perform a symbolic gesture associated with the object. In the example given, participants may use their index finger as if it were the toothbrush. This performance would be classed as “body part as a tool” error (BPT). Such errors (for a critical discussion see Cubelli & Della Sala, 1996) occur when patients use a part of their body as if it were the object: in the example above, the finger is used as a pretended toothbrush, representing its “physical properties.” Healthy participants also produce BPT errors (Duffy & Duffy, 1989) on command (under verbal, visual or tactile conditions) as well as on imitation (Mozaz, 1992).

Raymer et al. (1997) used more detailed instructions and asked participants to “imagine holding and using the tools just as they would if they had the actual tool” (p. 290). They showed that only apraxic patients committed BPT errors. BPT errors reflect the retrieval of the semantic knowledge about an object in the absence of the procedural information specifying the correct way of handling or using it (McDonald, Tate, & Rigby, 1994). As a consequence, a symbolic gesture would be produced rather than a pantomime. Some objects (i.e., scissors) are particularly sensitive of being described in symbolic, conventional ways, using parts of one’s body as a tool (Duffy & Duffy, 1989, Appendix pp. 234–235). In some cases the symbolic gestures are formally identical to the pantomimes (for instance the gesture of smoking a cigarette is akin to miming its use), masking possible difficulties the patients may have in pantomiming. However, most BPT errors produced by apraxics are novel gestures, generated by the patients to convey the object’s meaning and function. For instance, some patients use their index finger as if it were a pen. BPT errors may result from a “compensatory strategy” (Heilman & Rothi, 1985, p. 147) to overcome failures in pantomiming.

Pantomimes should be consistent with the characteristics of the objects, the use of which has to be mimed. The correct gesture has to be performed considering distance (i.e., the distance of the fingers from the table when miming how to use a pen), configuration (i.e., shape of hand on object) and orientation of the acting hand. Failure to properly consider these features results in spatial or postural errors that are frequently observed in apraxia (e.g., Mozaz, 1992; Roy, Black, Blair, & Dimeck, 1998; for a complete taxonomy see Leiguarda & Marsden, 2000, Table 2, p. 864).

To execute a pantomime it is necessary to re-produce the posture sustained when holding the real object. As a result the pantomime is often a novel, unfamiliar and creative gesture that capitalises on lexical and semantic representations. Hence, the integrity of the lexical route is necessary for a correct pantomime. However, if pantomimes constitute a special class of gestures (at the same time meaningful and novel), their production may call for additional cognitive mechanisms. Our working hypothesis is that to produce pantomimes the information about the objects function (stored in the action semantics) and the motor programme for the object use (in the output lexicon) ought to be integrated to generate a complex, new gesture. Working memory, considered as a workspace (Della Sala & Logie, 2002; Logie, 1996), i.e., as a system that allows us to temporarily interact with and mentally manipulate long-term stored representations, would be a suitable candidate to fulfil the role of integrating learned knowledge about objects and procedures on how to use them to plan novel actions. Hints about this role of working memory can be found in the literature (e.g., Logie & Della Sala, in press; Logie, Engelkamp, Dehn, & Rudkin, 2001; Toraldo, Reverberi, & Rumiati, 2001).

In the course of a group study aimed at validating a test battery for apraxia (Bartolo, 2002) we came across the case of a patient whose pattern of performances could hardly be accounted for by the current models of praxis. The aim of the present study is twofold, to report on the case of this patient who presented with isolated deficits of pantomimes within the class of meaningful gestures (Experiment 1), and to show that these deficits are associated with a specific working memory defect (Experiment 2).

Section snippets

Case history

VL, a retired haberdasher with seven years of formal education, was 66 when she had a stroke affecting the left hemisphere basal ganglia and external capsule (see CT scan in Fig. 2).

Soon after the stroke, she was examined with the B.A.D.A. (Miceli, Laudanna, Burani, & Capasso, 1994), an Italian battery of tests assessing language functions. She showed a severe agraphia and some difficulties in language production but good language comprehension. In particular, she made some errors in naming

Testing procedures and scoring

A battery was devised to assess the production of four types of gestures (pantomimes, transitive, intransitive, and meaningless) on two testing conditions (on command and on imitation) and with different input modalities. A series of pilot studies including 120 participants (Bartolo, 2002) allowed us to select fifteen objects, fifteen symbolic gestures, and fifteen meaningless gestures (see Appendix). The purposes of the pilot studies was to ensure that the objects were not eliciting BPT errors

Experiment 2: Assessment of working memory

To test the hypothesis that working memory is necessary to correctly carry out pantomimes, we assessed VL with a classic test of working memory, the dual-task. A key function of the assumed workspace would be that of integrating information simultaneously activated from different long-term memory sources with new environmental inputs. The dual-task typically taxes the co-ordination function of working memory (Baddeley, 1996). Hence, it is a suitable candidate to fulfil the purposes of our

General discussion

The findings from this study demonstrate that apraxic deficits limited to pantomime execution could be observed (Experiment 1), and that the patient showing such deficits also presented with working memory dysfunction (Experiment 2) in the absence of executive or any other neuropsychological disorders (Table 1).

The current models of praxis processing cannot account for the deficits of pantomimes shown by VL. A variation in these models will be proposed (Fig. 4).

In particular, we propose that

Acknowledgements

We thank Claudio Villa, the artist known for the stories of Tex Willer, who drew the vignettes for the task assessing the production of symbolic gestures on visual input.

References (64)

  • A.M Raymer et al.

    The significance of body part as tool errors in limb apraxia

    Brain and Cognition

    (1997)
  • A Toraldo et al.

    Critical dimensions affecting imitation performance of patients with ideomotor apraxia

    Cortex

    (2001)
  • L Wang et al.

    Pantomime, praxis, and aphasia

    Brain and Language

    (1992)
  • N Allamanno et al.

    Problem solving ability in ageing and dementia: normative data on a verbal test

    The Italian Journal of Neurologica Science

    (1997)
  • D Andrewes

    Neuropsychology: From theory to practice

    (2001)
  • C Arrigoni et al.

    Constructional apraxia and hemispheric locus of lesion

    Cortex

    (1964)
  • A.D Baddeley

    Working memory

    Quarterly Journal of Experimental Psychology

    (1996)
  • A.D Baddeley et al.

    The decline of working memory in Alzheimer’s disease: a longitudinal study

    Brain

    (1991)
  • A Baddeley et al.

    Testing central executive functioning with a pencil-and-paper test

  • A Baddeley et al.

    Dual-task performance in dysexecutive and non dysexecutive patients with a frontal lesion

    Neuropsychology

    (1997)
  • A Baddeley et al.

    Dementia and working memory

    Quarterly Journal of Experimental Psychology

    (1986)
  • B Barquero et al.

    Imagery constraints on quantitative and qualitative aspects of mental synthesis

    European Journal of Cognitive Psychology

    (1999)
  • Bartolo, A. (2002). Apraxia: New tests for the assessment of a cognitive model and the evaluation of the semantic...
  • A Basso et al.

    Raven’s Coloured Matrices: normative values in 305 adult normal controls

    Functional Neurology

    (1987)
  • A.L Benton et al.

    Impairment in facial recognition in patients with cerebral disease

    Cortex

    (1968)
  • L Buxbaum

    Ideomotor apraxia: a call to action

    Neurocase

    (2001)
  • G Cocchini et al.

    Concurrent performance of two memory tasks: Evidence for domain-specific working memory system

    Memory Cognition

    (2002)
  • M Coltheart et al.

    Models of reading aloud: Dual-route and parallel-distributed-processing approaches

    Psychological Review

    (1993)
  • R Cubelli et al.

    The legacy of automatic/voluntary dissociation in apraxia

    Neurocase

    (1996)
  • S Della Sala et al.

    Dual task paradigm. A means to examine the central executive

    Annals of the New York Academy of Sciences

    (1995)
  • S Della Sala et al.

    Neuropsychological impairments of visual and spatial working memory

  • E De Renzi et al.

    Apraxia

  • Cited by (92)

    View all citing articles on Scopus
    1

    Authors are listed in alphabetical order.

    2

    Current address: Dipartimento di Neuroscienze TCR, Universitá di Modena e Reggio Emilia, Modena, Italy.

    View full text