Apraxias

Abstract

Apraxia is the loss of the ability to carry out skilled, learned, purposive movements that cannot be explained by weakness, sensory loss, incoordination, or poor comprehension — a disorder of the knowledge for action rather than of the motor apparatus itself. This article traces the disorder from Hugo Liepmann's 1900 division into ideomotor, ideational, and limb-kinetic forms, through Norman Geschwind's disconnection account, to the modern cognitive model of limb praxis that treats gesture as a two-route system linking stored action representations to motor output. It examines the left parietal and premotor basis of praxis, the taxonomy of errors that distinguishes the subtypes, and the debate over whether tool use rests on stored manipulation knowledge or on general technical reasoning. Three interactive demonstrations explore the two-route model, the lesion-to-syndrome mapping, and the classification of gesture errors.

Keywords: apraxia, ideomotor apraxia, limb praxis

Apraxia — from the Greek for “without action” — is the inability to perform skilled, learned movements on demand despite intact strength, coordination, sensation, and understanding of what is being asked. A patient with limb apraxia has normal power in the hand and can move it freely, yet asked to wave goodbye, salute, or show how to use a toothbrush, produces a clumsy, mislocated, or unrecognizable gesture. The failure is not in the muscles, nor in the basic motor pathways, nor in comprehending the instruction; it is a specific breakdown of the system that translates the idea of a skilled act into the movement that performs it (Gross & Grossman, 2008; Leiguarda & Marsden, 2000).

The disorder is defined by a set of exclusions as much as by its positive signs. Hugo Liepmann established the modern concept between 1900 and 1908, showing in a landmark series of cases that the deficit dissociates from paralysis, that it is provoked most reliably by asking for movements out of their natural context, and that it follows damage to the left hemisphere regardless of which hand is tested (Liepmann, 1900). Because skilled action can break down at more than one stage, apraxia is not a single condition but a family of them: the classical scheme distinguishes ideomotor apraxia, in which the form of a well-conceived gesture is corrupted; ideational apraxia, in which the plan for a multi-step action with real objects falls apart; and limb-kinetic apraxia, a loss of the fine, fluid dexterity of individual movements (Wheaton & Hallett, 2007).

The National Library of Medicine files apraxia in its Medical Subject Headings as a psychomotor disorder — formally, the “loss of the ability to carry out purposeful movements resulting from disorders of the cerebral cortex” — and classifies it, in the disease trees, among the neurologic manifestations and the psychomotor disorders. It is, in other words, an acquired pathology of skilled action, and this article treats it as such, describing the injuries that produce it and the cognitive stages they disrupt.

Types of Apraxias

In the Medical Subject Headings, apraxia sits beneath the broader heading of psychomotor disorders, and the classification enumerates three narrower descriptors directly under it (Table 1). MeSH is an indexing vocabulary built for retrieving the biomedical literature, not a clinical taxonomy, so this formal tree cross-cuts the scheme clinicians actually use: it files by named syndrome rather than by the ideomotor / ideational / limb-kinetic stages of the classical division, so a reader should not expect the two systems to line up. The three MeSH children are named here as the classification files them; none is yet a separate article on this site, so none is linked.

The clinically useful distinctions, however, are drawn along the axis Liepmann introduced: which stage of skilled action has failed — the plan for a sequence of acts, the spatiotemporal form of a single gesture, or the fine execution of the movement itself. It is that axis, developed in the sections below, that organizes the modern understanding of the limb apraxias and their assessment.

The Ideomotor–Ideational Distinction

Liepmann's enduring contribution was to show that skilled action is not a single step but a layered process, and that apraxia can strike different layers. In ideomotor apraxia — the most common and most studied form — the patient knows what to do and in what order, but the form of the movement is corrupted. Asked to pantomime using a hammer, they may orient the hand wrongly, place the action in the wrong plane, use the arm as if it were the hammer itself (a “body-part-as-object” error), or produce clumsy, poorly timed strokes. The knowledge of the act survives — the patient often recognizes a correct gesture when someone else performs it — but cannot be expressed cleanly in movement (Heilman et al., 1982; Buxbaum, 2001).

In ideational apraxia the deficit lies earlier, in the plan itself. Single gestures may be performed adequately, but a task requiring several actions in the correct order with real objects — filling and lighting a pipe, preparing a letter for mailing, making a cup of coffee — disintegrates: steps are omitted, misordered, or performed with the wrong object, as though the script for the sequence were lost. Because it degrades the use of actual tools in daily life, ideational apraxia is often the more functionally disabling of the two, even though ideomotor apraxia is the more frequently examined (Leiguarda & Marsden, 2000; Zadikoff & Lang, 2005).

A related but distinct deficit, conceptual apraxia, is a loss of the knowledge of tool function itself — which tool performs which action, and what mechanical advantage it confers — so that the patient may select or pantomime the wrong tool for a task even when the sequencing of steps is intact. Heilman and colleagues separated it from ideational apraxia to mark the difference between losing the content of tool knowledge and losing the organization of a multi-step plan, a distinction that maps onto the action-semantic component of the cognitive model developed below (Ochipa et al., 1992).

A third form, limb-kinetic apraxia, sits at the boundary with the elementary motor system: it is a loss of the fine, individuated, fluid movements of the fingers and hand — the deftness needed to pick up a coin or button a shirt — that cannot be attributed to weakness or to a corrupted action representation. Whether it belongs with the apraxias at all has been debated since Liepmann, precisely because it lies so close to the motor execution stage rather than to the knowledge of the act (Wheaton & Hallett, 2007; Zadikoff & Lang, 2005).

The Cognitive Model of Praxis

The clinical subtypes were given an information-processing spine by the cognitive neuropsychological model of limb praxis, which recast apraxia as damage to identifiable components of a gesture-production system rather than as a collection of syndromes. In this model, gesture is produced by two routes. A lexical (or indirect) route handles familiar, meaningful gestures: an input praxicon recognizes a seen gesture, an action-semantic system holds knowledge of what tools and acts are for, and an output praxicon stores the learned motor programs — the spatiotemporal “engrams” — that specify how each familiar act is performed. A direct (or sublexical) route bypasses stored knowledge to convert a seen movement straight into an imitated one, which is how a healthy person copies a novel, meaningless gesture they have never made before (Rothi et al., 1991; Cubelli et al., 2000).

The model's power is that it predicts dissociations. Damage to the output praxicon should impair the pantomime and imitation of familiar gestures while sparing the imitation of meaningless ones (the direct route is intact); damage to the direct route should do the reverse, sparing meaningful gestures while wrecking the imitation of novel ones. Both patterns are observed, and the double dissociation is strong evidence that the two routes are real and separable (Cubelli et al., 2000; Buxbaum, 2001). The first demonstration builds this two-route system, letting the reader choose a gesture type and a damaged component and watch which route carries the movement and whether it succeeds.

Note. Familiar meaningful gestures are produced through the lexical route (action semantics and the output praxicon); novel meaningless postures are copied through the direct route. Because the routes are separable, a lexical lesion spares imitation of novel postures and a direct-route lesion spares familiar gestures — the double dissociation that shows the two routes are real. Original schematic; outcomes are illustrative, not patient data.

That meaning itself modulates imitation — and is not merely a property of the lexical route — is shown by finger-posture studies in which apraxic patients imitate configurations better when the posture is given a meaning than when it is meaningless, implicating the semantic system in tasks once thought to run purely through the direct route (Achilles et al., 2016). The model has been refined rather than overturned: the strict separation of a purely visuomotor direct route from a meaning-based lexical one is now seen as a simplification, but the core claim — that gesture production draws on dissociable stored representations and a more direct imitative channel — remains the field's working framework (Buxbaum & Randerath, 2018).

The Neuroanatomy of Praxis

Liepmann's second great finding was anatomical: praxis is lateralized. Testing patients whose gestures failed regardless of which hand was used, he concluded that the left hemisphere houses the representations that guide skilled movement for both sides of the body, and that apraxia follows when those representations are damaged or cut off from the motor cortex that would execute them (Liepmann, 1900). Geschwind gave this a disconnection interpretation: a command to the left hemisphere reaches its language areas, but if the fibers carrying the plan forward to the motor cortex — or across the corpus callosum to the right motor cortex controlling the left hand — are severed, the intact motor system never receives the instruction, producing apraxia by disconnection rather than by destruction of a center (Geschwind, 1965).

The modern localization centers on the left inferior parietal lobe, which appears to store the spatiotemporal representations of skilled acts, working with left premotor and prefrontal regions that select and sequence them. Lesion and imaging studies converge on this parietal-premotor praxis network: damage to the left parietal cortex impairs the knowledge of how gestures should look and unfold, while more anterior damage disrupts their selection and control, and the pantomime of tool use in particular depends on left parietal integrity (Goldenberg, 2009; Niessen et al., 2014). The reframing of apraxia as a disorder of higher-order sensorimotor integration — the transformation of stored knowledge into the spatial and temporal parameters a movement needs — captures why parietal damage is so central: it is there that knowledge for action is turned into a form the motor system can run (Leiguarda & Marsden, 2000). The second demonstration lets the reader place a lesion at different sites in this network and see which apraxic syndrome results.

Note. The stored spatiotemporal gesture representations are destroyed, so pantomime and imitation are distorted in both hands, with spatial errors. The right hemisphere and its motor cortex are not drawn; the callosal case is apraxic on the left because that hemisphere is cut off from the left-hemisphere representations. Original schematic; a simplified lateral view, not an anatomical atlas.

Assessment and Error Types

Because apraxia is defined by exclusion, its assessment is a structured search for a movement deficit that survives after weakness, sensory loss, incoordination, and poor comprehension have been ruled out. The examination samples gesture across three conditions that the cognitive model predicts will dissociate: pantomime to command (“demonstrate how a key is used”), which taxes the whole lexical route and is the most sensitive to ideomotor apraxia; imitation of the examiner's gestures, both meaningful and meaningless, which separates the lexical and direct routes; and actual tool use with the real object in hand, which is often the least impaired because the tool itself supplies cues that the stored program need not (Dovern et al., 2012; Gross & Grossman, 2008).

What is scored is not merely success or failure but the kind of error, because the error taxonomy is what distinguishes the subtypes and localizes the deficit. Spatial errors corrupt the posture, orientation, or plane of the movement, or substitute the body part for the tool; temporal errors disturb the timing, sequencing, or rhythm of the strokes; and content errors produce the wrong gesture altogether, such as pantomiming the use of a different tool. Spatial and temporal errors are the signature of ideomotor apraxia, whereas omissions and misorderings of steps mark the ideational form (Wheaton & Hallett, 2007; Buxbaum & Randerath, 2018). The third demonstration presents a target gesture and lets the reader introduce each error type in turn, seeing how the resulting movement would be classified.

Note. Bedside scoring records the type of error, not merely pass or fail. Spatial errors (wrong posture, plane, or the limb used as the tool) and temporal errors (disturbed timing and rhythm) are the signature of ideomotor apraxia; a content error produces the wrong gesture entirely. Original schematic; a single static pose per error, not motion-captured data.

Worked Example

The subtypes of apraxia are best distinguished by working through the pattern of a bedside examination, because the diagnosis turns on how performance varies across tasks rather than on any single failure. Consider a patient recovering from a left-hemisphere stroke who, asked to pantomime using a hammer, produces a recognizable but distorted gesture: the hand is held flat instead of gripping, the striking action is aimed sideways rather than down, and after a few clumsy strokes the fist itself becomes the hammer. Strength, sensation, and comprehension test normally, and the patient nods when shown a correct pantomime, agreeing it is right.

That single observation already localizes the failure. The preserved comprehension and recognition place the deficit after the plan and the action-semantic system; the spatial errors — wrong posture, wrong plane, body-part-as-object — are the classic signature of a corrupted gesture form. This is ideomotor apraxia. The diagnosis is confirmed by two further contrasts. First, imitation: asked to copy the examiner's hammering gesture, the patient improves little, because the same output representation feeds both pantomime and the imitation of a meaningful act — yet asked to copy a meaningless hand posture, they may do better, because that runs through the spared direct route. Second, actual tool use: handed a real hammer and a nail, the patient performs far more competently than in pantomime, because the tool in the hand supplies the cues the degraded stored program cannot.

Now contrast a second patient who pantomimes and imitates single gestures well but, given a candle, a matchbox, and asked to light the candle, strikes the unlit match against the candle, then tries to light the box: each elementary act is intact, but the sequence has fallen apart. Here the plan itself is disrupted — ideational apraxia — and the contrast with the first patient makes the logic explicit. Spatial and temporal distortion of a well-ordered gesture points to the form stage (ideomotor); disorganization of a correctly formed multi-step action with objects points to the plan stage (ideational). The same branching logic — pantomime versus imitation versus real tool use, scored by error type — is what the three demonstrations animate, and what turns a vague “clumsiness” into a localized diagnosis.

Discussion

Apraxia matters to cognitive psychology far beyond the clinic because it dissects skilled action into parts that ordinarily fuse into a single fluent act. Liepmann's division of praxis into a plan, a form, and an execution — drawn from lesioned patients at the turn of the twentieth century — anticipated the information-processing decomposition that the cognitive model later made explicit, and the two frameworks now reinforce each other: the ideomotor–ideational distinction is, in large part, a distinction about which component of the gesture-production system has failed (Liepmann, 1900; Rothi et al., 1991). That skilled action is left-lateralized, and that its representations sit in parietal cortex rather than in the motor strip, tells us the brain treats the knowledge of how to act as a form of cognition distinct from the machinery of movement (Goldenberg, 2009).

The disorder also anchors a broader argument about the mind's architecture. The double dissociation between imitating meaningful and meaningless gestures, like the sparing of tool use relative to pantomime, shows that action is served by multiple, partly independent routes rather than a single pipeline — the same fractionation logic that recognition disorders such as agnosia reveal on the perceptual side (Cubelli et al., 2000). And because apraxia is common after left-hemisphere stroke and predicts difficulty with everyday tasks, it has become a target for rehabilitation as well as a theoretical testbed, so that clarifying its cognitive structure has direct consequences for how the disorder is treated (Dovern et al., 2012).

Current Directions

The liveliest current debate concerns the very nature of the knowledge that apraxia damages: is tool use guided by stored sensorimotor programs, or by reasoning about objects in the moment? The traditional manipulation-based view, on which the cognitive model rests, holds that using a familiar tool retrieves a learned gesture engram; the competing reasoning-based account argues that people instead infer how to act on an object from its physical properties — a form of technical reasoning about mechanical possibilities — so that apraxia reflects impaired reasoning rather than lost motor memories. The two accounts make different predictions about which patients should fail which tasks, and adjudicating them is reshaping how tool-use deficits are understood (Osiurak & Badets, 2016; Osiurak et al., 2017).

A second, connected line of work reexamines the boundary between the model's two routes. Evidence that giving a posture meaning improves apraxic imitation implies that semantics penetrates tasks once assigned to the purely visuomotor direct route, blurring the strict dual-route separation and prompting more graded, interactive models of gesture production (Achilles et al., 2016; Buxbaum & Randerath, 2018). At the same time, finer analyses of pantomime — a task that is neither ordinary tool use nor pure imitation but a communicative act in its own right — are clarifying why it is so sensitive to left parietal damage and what its several components actually demand (Goldenberg, 2017; Niessen et al., 2014). Together these directions are moving the field from a fixed box-and-arrow diagram toward a dynamic account of how knowledge, perception, and reasoning combine to produce a skilled act.

Common Misconceptions

Apraxia is a kind of weakness or paralysis.

No. Strength, tone, and coordination are intact; the same patient who cannot pantomime hammering on command may hammer competently with a real hammer. The deficit is in translating the idea of a skilled act into movement, not in the muscles that would carry it out (Gross & Grossman, 2008).

A patient with apraxia does not understand the instruction.

Comprehension is tested and preserved before apraxia is diagnosed. Ideomotor apraxics typically recognize a correct gesture when they see one, showing the knowledge of the act is intact even when its production is not (Heilman et al., 1982).

Apraxia is a rare curiosity with little everyday impact.

It is common after left-hemisphere stroke and predicts real difficulty with daily activities such as dressing and using utensils, which is why it is a target of rehabilitation, not merely a diagnostic sign (Dovern et al., 2012).

Glossary

Action semantics.

Stored conceptual knowledge of what tools and actions are for, drawn on by the lexical route when a familiar gesture is produced from memory.

Apraxia.

The loss of the ability to carry out skilled, learned, purposive movements despite intact strength, sensation, coordination, and comprehension.

Body-part-as-object error.

A spatial pantomime error in which the limb itself is used as the tool — a finger for a toothbrush, the fist for a hammer — instead of representing the hand holding it.

Conceptual apraxia.

A loss of the knowledge of tool function — which tool performs which action and what it is for — distinct from the sequencing failure of ideational apraxia; produces content errors such as selecting the wrong tool.

Direct route.

The sublexical pathway that converts a seen movement straight into an imitated one, supporting the copying of novel, meaningless gestures without recourse to stored knowledge.

Disconnection syndrome.

A deficit produced not by destroying a cortical center but by severing the fiber tracts that carry information between centers; Geschwind's account of apraxia as the plan being cut off from the motor cortex.

Ideational apraxia.

A failure of the plan for a multi-step action with real objects, so that individual acts survive but their sequencing disintegrates.

Ideomotor apraxia.

A corruption of the spatial and temporal form of a well-conceived gesture, most evident when a familiar act is pantomimed to command.

Imitation.

The copying of an observed gesture; a core assessment task because meaningful and meaningless imitation dissociate the lexical and direct routes.

Lexical route.

The indirect pathway that produces familiar, meaningful gestures by way of the input praxicon, action semantics, and the output praxicon.

Limb-kinetic apraxia.

A loss of the fine, individuated dexterity of the fingers and hand that cannot be attributed to weakness or to a corrupted action representation.

Pantomime.

The gesturing of a tool's use without the tool in hand, on command or to a named object; the most sensitive task for ideomotor apraxia.

Praxicon.

The store of learned gesture representations; an input praxicon recognizes seen gestures, and an output praxicon holds the motor programs that produce them.

Praxis.

The capacity for skilled, learned, purposive movement; apraxia is its impairment.

Technical reasoning.

The proposed inference of how to act on an object from its physical properties, offered as an alternative to stored motor programs in accounting for tool use.

Key Researchers

Laurel J. Buxbaum (contemporary). Cognitive neuroscientist at the Moss Rehabilitation Research Institute and Thomas Jefferson University; advanced the action-representation account of ideomotor apraxia and its parietal basis. Google Scholar

Norman Geschwind (1926-1984). American behavioral neurologist who reinterpreted apraxia as a disconnection syndrome, in which the plan for action is severed from the motor cortex that would execute it. Wikipedia - Wikidata

Georg Goldenberg (contemporary). Neurologist and neuropsychologist at the Technical University of Munich; a leading theorist of apraxia and the parietal basis of gesture, imitation, and pantomime. Google Scholar

Kenneth M. Heilman (1938-2024). American behavioral neurologist at the University of Florida who distinguished the two forms of ideomotor apraxia and developed the representational account of praxis. Wikipedia - Wikidata - Google Scholar

Hugo Liepmann (1863-1925). German neurologist who founded the modern study of apraxia, dividing it into ideomotor, ideational, and limb-kinetic forms and establishing that praxis is a left-hemisphere function. Wikipedia - Wikidata

Francois Osiurak (contemporary). Professor at the Universite Lumiere Lyon 2; principal proponent of the reasoning-based account of tool use, a leading contemporary alternative to manipulation-based views of apraxia. ORCID - Google Scholar

Sergio Della Sala (contemporary). Professor at the University of Edinburgh; co-author of the cognition-in-action test that operationalized the cognitive model of limb apraxia. Wikipedia - Wikidata - Google Scholar

Peter H. Weiss (contemporary). Neurologist at the University of Cologne and the Julich Research Centre; studies the diagnosis, neural basis, and rehabilitation of upper-limb apraxia. ORCID - Google Scholar

Frequently Asked Questions

What is apraxia?
Apraxia is the loss of the ability to perform skilled, learned movements on command — such as waving, saluting, or showing how to use a tool — even though strength, sensation, coordination, and understanding of the task are all intact (Gross & Grossman, 2008).

How is apraxia different from weakness or paralysis?
Weakness is a failure of the motor apparatus; apraxia is a failure of the knowledge that guides it. An apraxic patient has normal power and may use a real tool competently, yet cannot pantomime its use on command, showing the problem is not in the muscles (Leiguarda & Marsden, 2000).

What is the difference between ideomotor and ideational apraxia?
Ideomotor apraxia corrupts the form of a single gesture — wrong posture, orientation, or timing — while ideational apraxia disrupts the plan for a multi-step action with real objects, so the steps are omitted or misordered (Wheaton & Hallett, 2007).

Which part of the brain is damaged in apraxia?
Praxis is chiefly a left-hemisphere function, and the left inferior parietal lobe — working with premotor and prefrontal regions — stores and sequences the representations of skilled acts; damage there, or its disconnection from motor cortex, produces apraxia (Goldenberg, 2009; Geschwind, 1965).

Why can an apraxic patient use a real tool but not pantomime it?
The real tool supplies perceptual and mechanical cues that the degraded stored motor program no longer provides, so actual tool use is often far better preserved than pantomime to command (Dovern et al., 2012).

How is apraxia tested?
By sampling gesture across pantomime to command, imitation of meaningful and meaningless gestures, and actual tool use, and by scoring the type of error — spatial, temporal, or content — after weakness, sensory loss, and poor comprehension have been excluded (Dovern et al., 2012).

Is apraxia the same as clumsiness?
No. Everyday clumsiness reflects the motor system; apraxia is a specific, lateralized disorder of learned action with a characteristic error pattern, diagnosed only after elementary motor and sensory causes are ruled out (Buxbaum, 2001).

Can apraxia be treated?
Rehabilitation can help. Because apraxia is common after left-hemisphere stroke and interferes with daily activities, gesture training and strategy-based therapies are used to improve function, and clarifying the disorder's cognitive structure guides that treatment (Dovern et al., 2012).

References