Clinical manifestations

Presentation and course

How apraxia is manifested clinically is frequently debated. De Renzi stated that ideomotor apraxia is an enduring syndrome that has no functional clinical effects (29). It has been generally assumed that environmental cues provided by the real contexts improve producing skilled movements. Alternatively, it has been suggested that cognitive demands may differ in natural contexts (where movements are guided by the physical conditions) and gesture imitation and pantomime (requiring the support of memory or problem-solving ability) (131). Thus, apraxia, when tested on order from the examiner, may essentially reflect a disorder of performance for an observer, ie, acting or obeying the director’s orders under contrived circumstances such as in a play, which brings increased attention to the action. A possible instance of the effect of varying attention toward acting upon a specific object during formal testing was suggested by the study by Neiman and colleagues (91). They found that movement errors were greater in brain-impaired patients when they acted upon a single object (eg, applying a postage stamp) than when acting upon a series of objects related to each other (eg, lighting a candle when given the candle, candle holder, and a match to manipulate). In a similar manner, a patient who falls below expectation to vocalize with a directed specific emotion (eg, anger, sadness) may actually fail because of unfamiliarity with the circumstance of contrived, laboratory-based behavior testing rather than the inability to voice appropriate emotions regardless of circumstances, which is termed aprosodia (04).

Noteworthy, apraxic patients are also impaired on aiming movements without visual feedback and on single finger tapping (62), thus not involving strictly skilled movements. Furthermore, apraxic patients make significant direction errors and have poor coordination of shoulder and elbow torques when starting a movement, consistent with defects in movement planning (90). Consequently, such observations may reflect a general disorder of coordinating the upper extremity during reaching, unrelated to tool use. Therefore, by definition the degrading of movement during actual or pretended functional movements would not be considered as apraxia because the movement disorder was not specific to skillful movements. In a related study, apraxic patients, nonapraxic brain impaired patients, and healthy individuals all showed incoordinated movements on formal apraxia tests (67). Thus, incoordination is not specific to apraxia. As a result, confusion reigns with diagnosing apraxia because clinical investigators have not specified what kinds of movement disorders should be considered as “basic” or otherwise.

An unresolved problem is that because apraxia most often follows from left hemisphere damage and thus often causes dominant hand paresis, this commonly requires the patient to rely upon the nondominant left hand for accomplishing everyday activities. Although at first glance, when lateralized brain hemispheric damage causes contralateral hemiparesis, it is commonly assumed that the other upper limb is not affected. However, detailed examination shows that the “nonparetic” limb is slowed on fine, dexterous activities (eg, using a feeding utensil, stacking checkers) (07). Consequently, it is unclear whether research on apraxia can identify hemiparetic patients who have truly unaffected contralateral limbs with regard to basic motor control.

An alternative manifestation of apraxia, hypothetically, would be the inability to remember how to conduct a functional action despite retaining the capability of unskilled movements (48). Here the literature has little evidence convincingly in support of this etiology of apraxia with actual object use. In one case, a stroke patient, after returning home, found himself unable to remember how to use specific tools such as an awl, and he could not write (14). The patient also had a general amnesia, eg, forgetting errands. Thus, the patient’s forgetfulness did not apply strictly to tool use. Although the patient did not have a disorder either of planning or sequencing functional movements, the presence of a general cognitive disturbance did not satisfy a definition of apraxia.

Similar instances have been published, which indicated impaired recall of how to use tools following corticobasal syndrome. A surgeon complained of forgetting how to perform specific intraoperative procedures (113). A dentist had to retire because he complained that he could not recall how to orient the handheld tools in his profession or to use pliers at home (87). However, because these studies did not include general cognitive assessments, it was not clear whether either a general memory or a spatial disorder could have contributed to the deficit.

A special instance is apraxic writing, also termed apraxic or apractic agraphia. Unlike virtually all other activities, writing leaves an enduring, visible trace of the object’s motion (using pen, etc.) that can be critically evaluated by the examiner at leisure. Despite being able to accurately copy written examples and correctly spell aloud, an apraxic agraphic patient may be unable to write intelligibly without a model of writing or print in front for viewing. The traces by such patients are either illegible or show considerably deformed but still recognizable letters or ideograms (eg, symbols used in Chinese writing) appropriate for one’s culture, particularly when compared to the patient’s writing samples prior to the brain illness. However, a brain illness that results in loss of movement ability by the preferred hand for writing, thus forcing the patient to rely on the previously nonpreferred hand, can challenge deciding whether the resulting distortions of writing should be truly considered apraxic agraphia. Other forms of agraphia can also occur after brain illness. Thus, aphasic agraphia involves normal letter formation but incorrect letter selection. Spatial agraphia involves either failing writing symmetrically across the page or diverging from horizontal or vertical orientation of writing that would be appropriate to one’s culture.

A right-handed patient was reported who suffered from a superior left internal capsule stroke, resulting in moderate right hemiparesis, and notably, normal performance on a nongraphic apraxia test but marked apraxic agraphia (25). The irregularities of writing also extended to producing ideograms (eg, dollar signs), less so with drawing pictures but still present. The authors concluded that this patient had “pure apraxic agraphia.” Although the authors judged that the patient had normal performance on the apraxia test, the unavailability of kinematic examination of the hand during apraxia testing might fail to detect degraded motor control from direct observation. Kinematic analysis of apraxic patients on simulated functional activities has shown objectively-measured irregular movement control compared to healthy individuals (24; 103; 104). Thus, the authors’ judgment of “pure” apraxic agraphia might have been caused by the general loss of fine motor control, which allowed a “pass” on the apraxia test due to not requiring as much motor precision on the associated tasks than on writing.

In one case, a 69-year-old man acutely developed the inability to carry out skilled movements with his right arm, instead producing awkward movements (70). He had had a parenchymal cerebral hemorrhage five years earlier. Further work-up demonstrated nonconvulsive status epilepticus on the electroencephalogram, involving continuous rhythmic theta-delta activity with a fluctuating frequency between 1.5 and 5 Hz in the left centro-parietotemporal area, without epileptic waveforms. The phenomena were abruptly resolved by intravenous levetiracetam and clonazepam. This suggests that nonconvulsive status epilepticus should be considered in the differential diagnosis of the acute presentation of loss of skillful movement.

Although ideomotor apraxia is usually observed in the upper limbs, apraxia can also be found in the lower limbs, particularly following large left hemisphere lesions (05). This uncommon disturbance was shown by asking patients to slide the nonparetic leg forwards and backwards, kick, cross the legs, pretend to stamp out a cigarette, and trace a cross with the foot.

Evidence for real-life ideational (or conceptual) apraxia appears more secure. Ochipa and colleagues reported the dramatic effects of conceptual apraxia on a stroke patient’s inability to select and manipulate appropriate real tools in everyday activities such as brushing teeth and eating a meal (92). Similar errors appeared in a traumatic brain injury patient (120). Examples to watch for in patients during spontaneous functional activities include:

Prognosis and complications

Complications of apraxia are not medical but instead functional. Thus, apraxia could affect the person’s ability to function in everyday life, especially with tool use and knowledge of object-tool-action relationships.

However, natural contexts usually provide cues for the automatic retrieval of the actions, whereas clinical testing represents an artificial condition that makes the movements more deliberate and less automatic (133). It is well known that in clinical testing conditions, ideomotor apraxia is much more evident, whereas in natural environmental conditions apraxia may remain relatively unnoticed. Rapcsak and colleagues studied a right-handed man with an extensive infarction in the left hemisphere associated with chronic right hemiplegia and Broca aphasia (108). Despite his severely impaired pantomiming of transitive gestures with his left hand, his actual object use and intransitive gestures were relatively spared. Moreover, this patient successfully lived alone, made home repairs, and even pursued his hobby of building model airplanes—all by using only his nondominant hand! This suggests that current laboratory and clinical apraxia tests may have limited value to predict real-world behaviors.

Several studies have suggested that ideomotor apraxia on within-clinic or laboratory assessments are a biomarker for dependency for everyday activities at the patient group level (37; 53; 123; 13; 66). Contrary findings were reported in one study (98). However, studies thus far have not evaluated ideomotor apraxia during everyday activities in the home. Furthermore, many case reports have shown that brain-illness patients who had failed on apraxia assessments have completed normal tasks in real-world circumstances (09; 45; 102; 30; 11; 55; 130). As a result, the ability to predict apraxia on real-world activities within individuals has not so far been established based on present apraxia assessments.

Few studies have reported long-term follow-ups of apraxia. Using different periods of assessment, several studies indicate that apraxia generally recovers after stroke after several months (08; 89; 68). Buccofacial apraxia has worse prognosis than limb apraxia.

Clinical vignette

A 55-year-old woman presented a cerebrovascular accident involving the parietal branches of the left middle cerebral artery. One week later in a neuropsychological evaluation she was found alert, cooperative, but time-disoriented. No right hemiparesis was observed, but although she was right-handed, it was evident that she preferred to use the left hand in diverse activities, such as pointing and taking objects. Her speech was moderately slow, but no significant changes in language were documented. Reading was normal, but holding the pencil with the right hand for writing was impossible. Using her left hand, she could only write the letter A. No hypoesthesia was found and she reported that tactile sensations were similar in her right and left hand. Two-point discrimination test did not demonstrate any significant difference in her ability to distinguish two different tactile stimuli placed in her right and left arm. Her ability to localize stimuli was also similar in both arms. She could recognize the fingers touched by the examiner in the right and in the left hand.

Performance on an ideomotor praxis ability test was abnormal. She was unable to mimic any of 10 movements by verbal command (eg, show me how to use the scissors). Parapraxias, spatial confusions, and movement simplifications were recorded. By imitation, performance improved. Difficulties were about the same in the right as in the left hand. Imitating hand position presented by the examiner, especially when using the right hand, was abnormal. Similarly, imitating 10 face movements and performing face movement by verbal command (eg, show me how to use a straw) was abnormal. When observing the patient’s performance in different daily activities (eg, taking her breakfast), however, no ideational apraxia was evident.

Biological basis

Etiology and pathogenesis

The most frequent etiology for apraxia is stroke. Brain tumors and head trauma in principle may also cause ideomotor apraxia. However, the latter illnesses present only as callosal apraxia. Progressive apraxia may appear in corticobasal syndrome (72; 115; 111; 128), frontotemporal lobar degeneration (126), progressive supranuclear palsy (127), Huntington disease (59; 61), and Alzheimer disease (96; 114). Limb apraxia is also found in multiple sclerosis, as identified by the Test of Upper Limb Apraxia (TULIA) test (66), although the study did not specify whether patients were having difficulty with performing everyday activities, if any. A long history describing erroneous gesture production in response to command has been noted in persons with schizophrenia, though not recognized as apraxia until recent years (137). Nonconvulsive status epilepticus has been reported to cause acute unilateral impairment of skillful movement, coinciding with a contralateral chronic brain lesion from a cerebral hemorrhage (70).

Liepmann's proposals (75; 76; 74) were resurrected by Geschwind in 1965 (44) who discussed them in the light of new human and animal data. Geschwind supported the notion of the dominance of the left hemisphere for learned movement skill and of the existence of movement formulae or memories (stored representations). He also supported Liepmann's statements that apraxia resulted from lesions in the dominant parietal region but argued that the important area was the underlying white matter pathways, specifically the arcuate fasciculus and not the cortex.

The clinical research literature primarily associates apraxia with dominant hemisphere pathology (57). However, the method for diagnosing apraxia may question whether there is association with damage to one particular hemisphere. Thus, when apraxia is diagnosed based only on errors on movement imitation rather than pantomiming tool use, there is no clear bias toward either hemisphere (31; 51; 49). Consequently, the diagnosis of so-called crossed apraxia, rarely reported (84; 109), is a matter of the manner of testing. The lesion laterality for pantomiming was more distinctively associated with the left hemisphere (13). However, because patients with ideomotor apraxia generally do not show such deficits in real-world spontaneous actions (29), the lesion laterality question pertaining to apraxia concerns only behavior under contrived circumstances.

Everyday naturalistic actions involving complex devices can be impaired either by left or right hemisphere disease. Hartmann and colleagues compared the performance of left- and right-hemisphere-damaged patients in two naturalistic tasks: preparing coffee with a drip coffee maker and fixing a cassette recorder (54). Deficits were observed in both groups. They suggest that different cognitive impairments cause failure in right versus left hemisphere disease. Left hemisphere disease may impede retrieval of functional knowledge from semantic memory, resulting in ideomotor apraxia. According to the authors, the major deficit in patients with right hemisphere disease results from the demand to keep track of multistep actions.

Apraxia is often associated with aphasia. In one study, about half of the patients with brain damage suffered from comorbid aphasia and apraxia (139). This association, however, has been polemic. It could be either because (1) they represent two manifestations of a common underlying defect, or (2) this association is the result of the proximity of their neural substrate. Seemingly, both explanations can be partially true. Using voxel-based lesion symptom mapping, two areas of overlap between aphasia and apraxia have been identified: the left anterior temporal lobe and the left inferior parietal lobe. At the same time, there are also areas in the parietal and temporal lobes associated exclusively with apraxia or aphasia (48). Apraxia has also been significantly associated with caudate nucleus infarction (118).

Ideomotor apraxia is associated with damage to the inferior parietal lobe, supplementary motor area, the corpus callosum, and subcortical structures. The left parietal cortex has been proposed to participate in planning, executing, and suppressing praxis movements (140). Patients with tool use impairment have lesions involving mainly the supramarginal gyrus, whereas patients with errors in the preceding grasping movement of tools usually present lesions including the left frontal and parietal cortices, especially the inferior frontal gyrus and the angular gyrus (106). Furthermore, two different brain streams for praxis have been proposed, associated with two apraxia variants: damage to the dorso-dorsal stream (involved in movements based on physical object properties), including the posterior intraparietal sulcus and superior parietal lobule, is associated with imitation defects, whereas damage to the ventral stream (involved in action representations), including regions such as the supramarginal gyrus and the extreme capsule, is associated with pantomime deficits (60). Buxbaum and colleagues observed in a sample of 71 left hemisphere stroke patients that deficits in the posture component of the tool-related gesture tasks were associated with left posterior temporal gyrus damage, whereas impairments in the kinematic component of different gesture tasks were correlated with left inferior parietal and frontal pathologies (21).

Although ideomotor apraxia is found more often following left than right hemisphere, a significant association has been found in right-hemisphere stroke patients between apraxia and unilateral spatial neglect (34). This finding suggests that the dysregulation of skilled movements during formal assessment may be affected by impaired spatial movement planning.

Conceptual apraxia is caused by damage to the action semantics system (110). It is observed in patients with lesions to posterior language areas and patient with dementia, specifically dementia of the Alzheimer type (38). Patients with conceptual apraxia have difficulty associating actions with objects. They also have impairment in the mechanical knowledge needed to associate tools with the objects that receive their action, recognize the advantages of using a tool in a given situation, and create tools for a specific task. When pantomiming to command, patients with conceptual apraxia will make semantically related and unrelated content errors. For example, when asked “show me how to use a hammer,” the patients may make the movements for a screwdriver (related) or a pen (unrelated). Ebisch and colleagues used fMRI to investigate the neural systems supporting conceptual knowledge for proper object use in a sample of normal subjects (36). Activation was observed in the left parietal-temporal-occipital junction and inferior frontal, anterior dorsal premotor, and presupplementary motor areas, apparently representing the human neural system for conceptual knowledge of proper object use.

Schwartz and Buxbaum report that patients with traumatic brain injury may produce errors of action in everyday activities such as making coffee and brushing teeth (119). The types of errors produced by the patients with traumatic brain injury are similar to those produced by patients with unilateral stroke with the most common errors being subtask omission and sequence. The authors noted that the performance of the patients with traumatic brain injury varied greatly in that the same patient may make apraxic errors in one task but not in another. Conceptual apraxia is associated with damage to caudal parietal lobe, temporoparietal junction but is most frequently associated with degenerative dementia of the Alzheimer type (93).

Conduction apraxia has been recognized to follow deficient imitation of actions with relatively preserved pantomiming actions (94). This disorder has most often been associated with damage to the supramarginal gyrus or Wernicke area, but specific lesion sites are not known. The localization of buccofacial apraxia is not clear, but it has been observed in cases of left frontal opercular (101) and left ventral premotor cortex pathology (69).

Dissociation or disconnection apraxia is associated with damage to the corpus callosum that disconnects the movement formulae of the dominant hemisphere with motor signal output to the opposite hemisphere. Callosal apraxia is caused by lesions of the corpus callosum, which disconnects the dominant hemisphere for language (left) from the hemisphere that controls the left hand (right) (99; 12). Therefore, pantomime to command, imitation, and use of real objects will be performed correctly by the right hand and incorrectly by the left hand.

Progressive apraxia may be observed in patients with degenerative disease such as Alzheimer disease (42; 26), Pick disease, progressive supranuclear palsy (127), corticobasal syndrome (63; 87; 115; 19), Parkinson disease (121), and progressive aphasia (65). It is interesting that the agrammatic variant of primary progressive aphasia has been found to present more praxis deficits but fewer cognitive disturbances than the logopenic variant (01). In addition, Rapcsak and colleagues reported a patient with slowly progressive bilateral apraxia associated with parietal lobe degeneration of unspecific origin (107).

In corticobasal syndrome, different types of apraxia can be observed, including ideomotor apraxia and limb-kinetic apraxia, although buccofacial and oculomotor apraxia can be found as well (142); limb-kinetic apraxia is usually dominant (127). Denes concisely compared the apraxia profiles associated with progressive supranuclear palsy and corticobasal syndrome (32). Patients with corticobasal syndrome were more impaired overall and were significantly more impaired when using tools in natural contexts than patients with progressive supranuclear palsy. In addition, patients with corticobasal syndrome initially present with apraxia in one limb. For both types of patients, impairment in the production of simple unimanual and bimanual dexterity tasks are observed. Apraxia in corticobasal syndrome has been related to gray matter reduction in the inferior frontal and premotor cortices, parietal operculum, superotemporal gyrus, and hippocampus of the left hemisphere. Severity of limb apraxia has been found significantly associated with parietal atrophy and anisotropy reductions in the parietofrontal associative fibers, whereas limb-kinetic apraxia has been related with reduction of hand sensorimotor connecting fibers (17).

Buxbaum and colleagues reported a patient with progressive ideomotor apraxia who was profoundly impaired on all gesture tasks (to auditory command, visual cue, and imitation) as well as imitation of meaningless gestures and on tasks that required spatiomotor transformations of bodily position, yet performed near normal when holding the tool (20). They concluded that the praxis system is dynamic and interacts continually with body schema knowledge in the posterior parietal cortex.

Apraxia in left-handers has been rarely analyzed. Goldenberg selected 50 consecutive left-handed patients with unilateral hemisphere lesions (47). Three different tasks were used to evaluate apraxia: pantomime of tool use, actual tool use, and imitation of meaningless hand and finger postures. Three aphasic patients with pervasive apraxia caused by left hemisphere lesions were found (apraxia and handedness were dissociated); three patients with pervasive apraxia due to right brain lesions without aphasia were also found (apraxia and aphasia were dissociated). In general, dissociations from handedness and from aphasia were found for the different apraxia manifestations, although the frequency was different according to the specific type of apraxia. Noteworthy, abnormal pantomime and abnormal tool use was usually associated with aphasia; but defective imitation of hand postures—associated with hemi-neglect—was more frequent after right hemisphere damage.

Prevention

The occurrence of nonconvulsive status epilepticus causing unilateral arm apraxia can be aborted by intravenous antiepileptic medication (70).

Differential diagnosis

Confusing conditions

Comprehension disorders, movement disorders, and language disorders (aphasias) must be systematically evaluated before a diagnosis of apraxia is made. One must also rule out visual agnosia when patients select the wrong tool for the task. The patient must be able to name the tool or describe its function and purpose to establish that the patient has apraxia and not visual agnosia.

Associated or underlying disorders

Because apraxia is associated more often with left hemisphere than right hemisphere disease, aphasia and apraxia commonly co-occur (95). A subset of apraxia disturbance, the impaired imitation of meaningless gestures, is associated with unilateral spatial neglect following right hemisphere stroke (34).

Diagnostic workup

Several relatively standardized models for evaluating apraxia have been proposed (03; 143; 85; 82; 105; 135; 134; 13; 71; 132). The plethora of models have hampered the development of a gold standard for diagnosis (10). In general, examination for apraxia should include the following:

However, despite the plethora of apraxia assessment batteries, there is no gold standard. Consequently, there has been no consensus concerning the core test items of apraxia. Activities that are commonly assessed include unimanual tool use such as hammer, telephone, screwdriver, scissors, and toothbrush, which are widely familiar. However, if the intention is to determine what core evaluations of an apraxia battery can optimize the sensitivity of such a battery—either to discriminate between brain-illness patients and health individuals, or to develop proxy measures for actual, spontaneous activities of daily living—comprehensive research has not determined the assessments that would be ideal. Conventional apraxia assessments have not kept pace with technological advances in tools that are commonly used (eg, computers, microwave ovens, power wheelchairs) (83). Hence, at present the diverse apraxia batteries largely have been decided by the personal preferences of the individual examiners rather than based on theoretical models that can be tested. As noted above, apraxia batteries have reflected failing scores during ideal laboratory or clinic environments, and yet exceptional brain-illness patients diagnosed with apraxia nonetheless have been independent with many or all routine activities of daily living in the real world. Further research is needed to assess the sensitivity of apraxia assessments for real-world dependency for activities of daily living. Such work thus far has not been conducted.

Another limitation of apraxia batteries has been the limited development of methods for scoring and diagnosing apraxia. Typically, an apraxia test battery compares a brain illness population to a brain-healthy, functionally independent population. Scoring the items of the apraxia test battery depends on observers who watch the pantomime or imitation of the actions (or less often, observing spontaneous real activities discretely) to decide whether they are correct or incorrect (41). Commonly, such batteries have established inter-rater reliability. However, the “pass” versus “fail” score is typically determined by the total of individual actions that are judged to have been essentially “normal.” The problem is that the pass/fail criterion (typically the mean number of failed items, which is at least two standard deviations below the mean by healthy individuals) may be decided by the composition of the particular test sample of brain-illness and healthy individuals. Hence, a score of less than 15 out of 20 correct actions might diagnose “apraxia” on one scale, but such a battery might lead to a different cut-off level in a different battery, or it may be possible that two different sample populations evaluated on the same apraxia battery might yield different cut-off scores.

Still another limitation is the scope of actions that are assessed on apraxia batteries. Typically, these involve 1-handed actions. Surprising is that obligatory 2-handed actions are seldom evaluated, for example, demonstrating or pretending the opening of a window in a room, shoe tying, shirt buttoning, or (usually) juggling. Another possibility would be to evaluate the coordination of hand and foot (ie, bicycling or shifting gears in a manual transmission car).

As a result, there is presently no gold standard of apraxia assessment that can benefit clinical practice. Moreover, observing completing functional activities in a clinic or laboratory to command (including some approaches of testing apraxia) may yield considerably different results compared to self-reported accomplishing everyday activities (122; 125).

A possibly helpful way to redefine and diagnose apraxia that would be clinically relevant would be to include only those abnormalities of voluntary, purposive, and skilled extremity movement that are actually disabling in the real-world setting. Thus, although many brain-impaired individuals may fail to accomplish skillful movements normally under artificial, laboratory conditions according to nonstandardized criteria, such individuals at times successfully reattempt movements to be more effective on meaningful tasks, such as using a dinner utensil (138).

The DATE (dementia apraxia test) has been shown to distinguish frontotemporal dementia from Alzheimer disease, in that the latter patients were distinctly more impaired on imitating meaningless limb gestures and pantomiming object use (64). The basis for this difference is suggested from the greater likelihood of disturbed parietal lobe-based modeling of specific limb-spatial configurations and greater semantic memory deficit affecting pantomiming object use in Alzheimer disease. The differential diagnosis of frontotemporal dementia versus Alzheimer disease when using the DATE has the sensitivity of 74% and specificity of 93%. The DATE was shown to be sensitive to self-reported real-world disability in persons with Parkinson disease (112), with the complication that some DATE items were confounded by concurrent motor impairment. Further revisions of DATE for use for Parkinson disease are needed.

The TULIA battery (Test of Upper Limb Apraxia) has been developed to specifically assess gesture production (135). The TULIA has been applied to assess deficits in gesture production in schizophrenia, which has been reported to be prone to discrepancies between gesture content and concurrent speech content and lend to confusing understanding the intended meanings of patients (06). The TULIA also has been found to help identify impaired self-efficacy for everyday activities in patients following a stroke (116).

Management

The compelling rationale for rehabilitating ideomotor apraxia depends on whether one views this disorder as a basis for dependency for everyday activities. As observed by Cubelli and Della Sala, the finding of laboratory-based ideomotor apraxia does not clearly translate to dependency in the real-life circumstance (28). Even disturbed hand orientation or temporal patterns on functional-relevant activities of daily living seen in the laboratory may allow the patient to spontaneously correct movements as needed for the task at hand in the real life (15; 80; 138; 18; 73). A limitation for rehabilitating ideomotor apraxia comes from the frequent finding that patients themselves do not recognize that they have a difficulty, which is a form of anosognosia (117). In one study, laboratory-assessed apraxia did not predict disability on the spontaneous use of tools or other common objects following stroke (136). Similarly, reproducing nonsense limb movements in the laboratory did not predict the inaction (learned nonuse) of the more affected arm in stroke patients when performing spontaneous limb use tasks (22).

To date, there has been a paucity of research focusing on treatment or management of apraxia (23; 33; 40). A series of studies by Maher and colleagues reported improvements in apraxia following direct treatment, but the improvements were item specific and generalization was not observed (81). Compared to patients without apraxia, stroke patients with apraxia in general present a similar level of improvement when attending an inpatient rehabilitation program, but their level of independence is lower not only upon admission but also when discharged (141). A study by Smania and colleagues assessed the effectiveness of a rehabilitation program for patients with ideomotor apraxia (124). Thirteen patients were randomized to a study group and a control group. The study group received 35 treatment sessions in which they underwent a behavioral training program comprising of gesture-production exercises. Results indicated that the treatment group improved significantly in their performance in ideational and ideomotor apraxia tests, whereas the performance of the control group did not change. The Smania study indicates that treatment specific to the gesture deficits found in patients with ideomotor apraxia may be effective. In another report, Smania and colleagues randomly assigned 33 patients with left hemisphere stroke, limb apraxia, and aphasia to an apraxia or a control (aphasia) treatment group. Apraxia severity was negatively associated with independence in activities of daily life. Apraxia treatment had a positive impact on praxic functions and independence in activities of daily life (123). However, to date the emphasis in rehabilitation has been on compensating for the disorder and managing the environment of the apraxic patient rather than directly treating apraxia. A randomized clinical trial of moderately disabled stroke patients on gesture training versus education resulted in no differences between groups, and the changes were modest (02). The general small size of patient samples and failure to retain improvements after training prevent recommending a specific form of rehabilitation for apraxia (43).

For patients with ideational apraxia, the following recommendations are given by Ochipa and colleagues (92):

It has been reported that transcranial direct current stimulation delivered to the left posterior parietal lobe reduces the time required to plan and perform skilled movements in both patients with ideomotor apraxia and healthy controls; this observation opens up new possibilities in the rehabilitation of apraxia (16).

Outcomes

Whenever apraxia has been noted to improve following specific therapy, the improvements have not been retained over the long term (43).