Cerebral Palsy

Abstract

Cerebral palsy is the most common physical disability of childhood: a group of permanent disorders of movement and posture arising from a non-progressive disturbance of the developing brain. Although defined by its motor signs, its interest for cognitive psychology lies in what accompanies them. A single early injury rarely spares cognition, so cerebral palsy is characteristically attended by difficulties of attention, working memory, processing speed, and executive function, sometimes present even when general intelligence is preserved. Function is graded not by a single severity score but by separate ordinal systems for gross motor, manual, and communicative ability. That profile, and the way motor disability constrains the assessment of cognition itself, make cerebral palsy a demanding test case for how the mind is measured when the body will not cooperate.

Keywords: cerebral palsy, gross motor function, cognitive profile

Cerebral palsy is not a single disease but an umbrella term for a group of conditions that share a common shape: a permanent, non-progressive disturbance of the immature brain that expresses itself as a disorder of movement and posture and, very often, of cognition (Rosenbaum et al., 2007). The lesion itself does not worsen, but its consequences change as the child develops and demands more of the injured system, so the clinical picture evolves even though the underlying injury does not (Bax et al., 2005). What makes cerebral palsy a subject for cognitive psychology, rather than paediatric neurology alone, is that the same early insult that disrupts the motor system rarely leaves the rest of the brain untouched: attention, memory, language, and executive control are frequently affected, and measuring them in a child who cannot reliably point, speak, or sit is a genuine methodological problem (Stadskleiv, 2020).

What Cerebral Palsy Is

Cerebral palsy is defined, in the consensus statement that governs modern practice, as *a group of permanent disorders of the development of movement and posture, causing activity limitation, that are attributed to non-progressive disturbances that occurred in the developing fetal or infant brain* (Rosenbaum et al., 2007). Every clause of that definition does work. *A group of disorders* signals that cerebral palsy is a description, not a diagnosis of cause; *permanent* excludes transient delays; *non-progressive* distinguishes it from degenerative disease, in which function is lost because the lesion itself worsens; and *developing brain* fixes the timing to the early period when the brain is still being built (Bax et al., 2005). The motor disorder is often accompanied by disturbances of sensation, perception, cognition, communication, and behaviour, by epilepsy, and by secondary musculoskeletal problems — the definition names these explicitly, because in practice they are the rule (Rosenbaum et al., 2007).

The historical thread runs back to William John Little, the English orthopaedic surgeon who in 1862 first linked difficult or premature birth and perinatal asphyxia to a later pattern of spastic motor disability — the condition long known by the eponym *Little's disease* (Colver et al., 2014). Little's causal story, that birth asphyxia is the usual culprit, dominated for a century and still shapes public belief, but modern epidemiology has substantially revised it: most cerebral palsy in children born at term is *not* explained by intrapartum asphyxia, and the reviewed risk factors point instead to a diverse set of prenatal and perinatal contributors (McIntyre et al., 2013), while a large and growing share of cases trace to genetic variants and developmental processes set in motion long before delivery (MacLennan et al., 2015). Prematurity is the single largest risk factor, and the prevalence of cerebral palsy climbs steeply as gestational age falls (Oskoui et al., 2013); population registries that have tracked cerebral palsy across successive birth cohorts anchor the overall rate at roughly two per thousand and trace how its origins have shifted over time (Himmelmann et al., 2010).

Clinically, cerebral palsy is subdivided by the dominant type of motor disorder — spastic (by far the most common, marked by increased tone and exaggerated reflexes), dyskinetic (involuntary, fluctuating movements), and ataxic (poor coordination and balance) — and by its topography, that is, which limbs are affected (Graham et al., 2016). Diagnosis is clinical, resting on the history and neurological examination, supported increasingly by magnetic resonance imaging that can identify the pattern and timing of the underlying brain injury in most cases (Korzeniewski et al., 2018).

Classifying Function, Not Just Severity

The decisive conceptual advance in cerebral palsy was to stop asking 'how severe is it?' and start asking 'what can the person do?' — and to answer that question separately for different abilities. The Gross Motor Function Classification System (GMFCS) grades gross motor function, especially walking and mobility, into five ordinal levels, from Level I (walks without limitation) to Level V (transported in a wheelchair with severe limitations in head and trunk control) (Palisano et al., 1997). Its power is that the levels are defined by everyday performance rather than by clinical signs, are stable and reliable, and — crucially — are meaningful to families; an expanded and revised version extended the levels through adolescence (Palisano et al., 2008).

The insight generalised. Manual ability is graded by the Manual Ability Classification System (MACS), which classifies how children handle objects in daily activities across five analogous levels (Eliasson et al., 2006), and communication by the Communication Function Classification System (CFCS), which grades everyday sending and receiving of messages (Hidecker et al., 2011). The three systems are deliberately *orthogonal*: a child can walk independently yet struggle to manipulate objects, or move well but communicate poorly, so a single global severity score would blur exactly the distinctions that matter for support and prognosis. Describing a person as, say, 'GMFCS II, MACS I, CFCS III' captures a profile that one number never could, and this profile-based description is now the common language of both clinical care and research (Graham et al., 2016). Table 1 sets out the three systems side by side, showing the separate ability each grades and what its least- and most-limited levels describe.

Table 1

The Three Function Classification Systems in Cerebral Palsy

Cognition in Cerebral Palsy

Because cerebral palsy is defined by its motor signs, its cognitive dimension is easy to overlook — and historically was. Yet the same early injury that disrupts motor control frequently disturbs cognition, and the cognitive difficulties are often the more limiting in the long run (Bottcher, 2010). The pattern is not uniform. Some children with cerebral palsy have intellectual disability; many have measured general intelligence in the typical range but show *specific* deficits — most consistently in selective attention, working memory, processing speed, and executive function — that can undermine learning even when overall IQ is preserved (Bottcher, 2010).

There is a further, subtler problem that belongs squarely to cognitive psychology: measuring cognition in cerebral palsy is confounded by the very motor and communicative impairments that define the condition. Most standardised cognitive tests demand a spoken answer, a pointed response, or a manipulated block within a time limit — precisely the outputs that cerebral palsy compromises. A child with intact reasoning but no reliable speech or hand control can score as intellectually disabled on an instrument that cannot separate *knowing* the answer from *producing* it (Stadskleiv, 2020). This is a measurement-validity failure of the first order: the test's motor and verbal demands are construct-irrelevant to the cognition it means to assess, and unless it is adapted — with eye-gaze response, extended time, or accessible formats — it systematically underestimates the abilities of exactly the children who most need them recognised (Stadskleiv, 2020). The cognitive profile of cerebral palsy therefore cannot be read off a standard IQ score; it must be reconstructed with methods that route around the motor bottleneck.

The Developmental Trajectory

If the brain lesion is non-progressive, why does function change? The answer distinguishes cerebral palsy from both static and degenerative conditions. Gross motor function in cerebral palsy follows a developmental curve: children gain motor skills through early childhood, approaching a ceiling that depends on their GMFCS level, and the *height* of that ceiling is set early and predicts long-term function (Hanna et al., 2009). For children at the milder levels the ceiling is high and stable; for those at the more severe levels it is lower, reached sooner, and — importantly — may be followed by a measurable *decline* in gross motor function during adolescence, as growth, spasticity, contracture, and pain erode capacities that were previously present (Hanna et al., 2009).

This trajectory is why the injury is described as non-progressive while the disability is not static: the lesion does not worsen, but the growing body places rising biomechanical demands on an impaired system, and secondary problems accumulate (Bax et al., 2005). The clinical consequence is that cerebral palsy is a lifespan condition requiring anticipatory management rather than a fixed deficit to be rehabilitated once. The same logic applies to cognition: as academic and social demands escalate through schooling, a specific attentional or working-memory weakness that was invisible in early childhood can become the dominant limitation, so the cognitive profile, like the motor one, is a moving target that must be reassessed as development proceeds (Stadskleiv, 2020).

Worked Example

How much can we trust a GMFCS level? Because clinical decisions and research strata both rest on the classification, its *inter-rater reliability* — the degree to which two clinicians independently assign the same level — has to be quantified, and Cohen's kappa is the standard tool. The GMFCS was designed for exactly this and shows high reliability (Palisano et al., 1997); the calculation below shows what 'high' means and why it is not the same as raw agreement.

Suppose two therapists independently classify 100 children into GMFCS levels and agree on 85 of them. The raw observed agreement is:

$$p_o = \frac{85}{100} = 0.85$$

But some agreement is expected by chance alone, and kappa corrects for it. Suppose the marginal proportions with which each rater uses the five levels give an expected chance-agreement of:

$$p_e = \sum_{k=1}^{5} p_{1k}\,p_{2k} = 0.25$$

Then Cohen's kappa is the observed agreement in excess of chance, scaled by the maximum possible excess:

$$\kappa = \frac{p_o - p_e}{1 - p_e} = \frac{0.85 - 0.25}{1 - 0.25} = \frac{0.60}{0.75} = 0.80$$

A kappa of 0.80 sits at the boundary conventionally called 'substantial' to 'almost perfect,' and it is meaningfully lower than the 0.85 raw agreement precisely because it discounts the 25% of agreement the two raters would have reached by guessing with the same habits. The correction matters most when one or two levels are common: if most children in a sample are GMFCS I, raters will agree often just by both saying 'I,' and only kappa reveals how much genuine, chance-independent reliability remains (Palisano et al., 1997).

The same reasoning underwrites the classification's clinical value. A system with a kappa near 0.80 can be used to communicate a child's function between a physiotherapist, a paediatrician, and a school with confidence that they mean the same thing — the property that let the GMFCS become a shared language — whereas a subjective 'mild / moderate / severe' judgement, with no chance correction and no operational anchors, would not survive the same test (Palisano et al., 2008).

Discussion

The scientific interest of cerebral palsy for cognitive psychology is that it forces a separation between competence and performance that ordinary testing lets us ignore. In a typically developing child, the ability to reason and the ability to speak or point develop together, so a test that measures the second is a fair proxy for the first. Cerebral palsy severs that link: a child may possess the underlying cognitive competence and yet be unable to produce the motor or verbal performance the test requires (Stadskleiv, 2020). This makes the condition a standing reminder that our instruments measure outputs, not minds, and that the inference from one to the other depends on assumptions that a motor disorder can violate.

The three classification systems embody the constructive response to that problem. By grading gross motor, manual, and communicative function separately, and by anchoring each in observable everyday performance rather than a global impression, they resist the temptation to collapse a person into a single severity axis (Palisano et al., 1997). The demonstrations on this page track the resulting picture at three resolutions: the GMFCS explorer shows how function is graded into meaningful ordinal levels, the cognitive-profile display shows how specific abilities can dissociate from general intelligence, and the trajectory curve shows how function evolves across development even though the injury does not. Where the account remains incomplete is in linking the pattern of early brain injury, now increasingly visible on imaging and traceable to genetic and prenatal origins, to the specific cognitive profile that follows — the causal chain from a periventricular lesion to a working-memory deficit is not yet read in detail (Korzeniewski et al., 2018).

Cognitive Implications

For cognitive psychology, cerebral palsy is a natural experiment on the measurement of cognition under motor constraint. It sharpens a distinction that is usually left implicit — between construct-relevant and construct-irrelevant task demands — because in cerebral palsy the construct-irrelevant demands (speak, point, manipulate within a time limit) are exactly what the disorder attacks. Building valid cognitive assessment for these children means stripping a task down to the reasoning it intends to probe and re-routing the response around the motor bottleneck, whether by eye-gaze selection, accessible response formats, or relaxed timing (Stadskleiv, 2020). The exercise clarifies, for cognitive testing generally, how much of a 'cognitive' score is really a motor and linguistic one.

The condition also speaks to the psychology of attention and executive function. The specific deficits most consistently reported in cerebral palsy — in selective attention, working memory, and processing speed — are the components of the control system that coordinates goal-directed behaviour, and their vulnerability to early diffuse injury suggests these functions depend on distributed networks that a focal lesion can disrupt disproportionately (Bottcher, 2010). That specific abilities can be impaired while general intelligence is spared is itself informative: it supports a componential rather than a monolithic view of cognition, in which processing speed and working memory are separable resources that early injury can dissociate, and it argues for intervention aimed at the specific weak component rather than at a global 'cognitive delay' (Stadskleiv, 2020).

Current Directions

Three fronts are active. The first is causation and prevention: the old model centred on birth asphyxia has given way to a picture in which prenatal and genetic factors dominate, and sequencing studies now implicate rare and de novo variants in a meaningful fraction of cases, reframing part of cerebral palsy as a neurogenetic condition and opening genuinely preventive questions (MacLennan et al., 2015). Alongside genetics, the recognised contribution of intrapartum events continues to shrink as the antenatal origins of most term cerebral palsy become clearer (Korzeniewski et al., 2018).

The second is early, accurate diagnosis. It is now possible to identify cerebral palsy, or high risk of it, well before the traditional age, by combining neonatal neuroimaging with standardised assessments of general movements and neurological examination — a shift that matters because the developing brain's plasticity is greatest early, so intervention begun in infancy may do more than the same intervention begun at three or four (Novak et al., 2017). The third front is evidence-based intervention itself: systematic reviews that grade every available therapy by the strength of its evidence — the 'traffic lights' syntheses — have imposed discipline on a field once crowded with untested treatments, distinguishing the interventions that work from those that merely persist (Novak et al., 2013), (Novak et al., 2020). Across all three the cognitive dimension is increasingly treated as a primary outcome rather than an afterthought, in keeping with the recognition that cognition often limits participation more than mobility does (Stadskleiv, 2020).

Common Misconceptions

Cerebral palsy is a progressive disease that gets worse over time.

No. The underlying brain injury is non-progressive — it does not worsen. What can change is function, as a growing body places rising demands on an impaired system and secondary problems such as contracture accumulate; the lesion is static even when the disability is not (Bax et al., 2005).

Cerebral palsy always means intellectual disability.

No. Cognitive outcomes vary widely. Many people with cerebral palsy have general intelligence in the typical range; some have specific deficits with preserved IQ, and motor impairment can cause cognition to be underestimated by tests that demand speech or hand use (Stadskleiv, 2020).

Cerebral palsy is caused by oxygen deprivation at birth.

Usually not. Intrapartum asphyxia explains only a minority of cases, especially in children born at term; most cerebral palsy traces to prenatal, developmental, and genetic factors, with prematurity the largest single risk (MacLennan et al., 2015).

A single severity score captures how affected someone is.

No. Gross motor, manual, and communicative functions are graded separately and can dissociate, so a person is described by a profile across the GMFCS, MACS, and CFCS rather than by one number (Graham et al., 2016).

Glossary

Ataxic cerebral palsy.

A type marked by poor coordination and balance, reflecting involvement of cerebellar pathways; the least common of the main motor types.

Cerebral palsy.

A group of permanent disorders of movement and posture attributed to a non-progressive disturbance of the developing fetal or infant brain.

Cohen's kappa.

A statistic measuring inter-rater agreement corrected for the agreement expected by chance; used to quantify the reliability of function classifications.

Communication Function Classification System (CFCS).

A five-level ordinal system grading everyday effectiveness of sending and receiving messages, independent of the motor and manual systems.

Construct-irrelevant demand.

A task requirement, such as speaking or pointing, that is not part of the ability being measured but which motor impairment can turn into a barrier, biasing the score.

Dyskinetic cerebral palsy.

A type marked by involuntary, fluctuating movements and variable muscle tone, reflecting involvement of the basal ganglia.

Gross Motor Function Classification System (GMFCS).

A five-level ordinal system grading gross motor function and mobility by everyday performance, from independent walking (I) to severely limited self-mobility (V).

Little's disease.

The historical name for spastic cerebral palsy, after William John Little, who first linked difficult birth to later motor disability in 1862.

Manual Ability Classification System (MACS).

A five-level ordinal system grading how children handle objects in daily activities, independent of gross motor function.

Non-progressive.

Describing an injury that does not worsen over time; in cerebral palsy the lesion is fixed even though its functional consequences can change with development.

Perinatal.

Relating to the period shortly before and after birth; one of several windows during which the disturbance underlying cerebral palsy can occur.

Prevalence.

The proportion of a population with a condition at a given time; for cerebral palsy, about 2.1 per 1,000 live births overall, rising steeply with prematurity.

Spasticity.

A velocity-dependent increase in muscle tone with exaggerated reflexes; the hallmark of spastic cerebral palsy, the most common type.

Topography.

The distribution of motor involvement across the body — for example hemiplegia (one side), diplegia (legs more than arms), or quadriplegia (all four limbs).

Key Researchers

Nadia Badawi

Macquarie Group Foundation Chair of Cerebral Palsy at the University of Sydney and the Cerebral Palsy Alliance Research Institute, whose epidemiological work clarified the risk factors for cerebral palsy in children born at term. Faculty Page

Ann-Christin Eliasson

Professor of occupational therapy at the Karolinska Institutet, developer of the Manual Ability Classification System (MACS) for grading how children with cerebral palsy handle objects. Faculty Page

William John Little

(1810–1894). English orthopaedic surgeon whose 1862 lecture first connected difficult birth and perinatal asphyxia to later spastic motor disability, giving the historical eponym *Little's disease*. Wikipedia

Alastair MacLennan

Professor at the University of Adelaide and leader of the Australian Collaborative Cerebral Palsy Research Group, whose work established the substantial role of genetic and prenatal factors over intrapartum causes. ORCID

Iona Novak

Professor at the University of Sydney and the Cerebral Palsy Alliance Research Institute, whose systematic reviews graded the evidence for cerebral palsy interventions and advanced early, accurate diagnosis. ORCID

Robert Palisano

Professor at Drexel University and lead developer of the Gross Motor Function Classification System (GMFCS) and its expanded and revised edition. ORCID

Peter Rosenbaum

Professor of paediatrics at McMaster University and co-founder of CanChild, central to the consensus definition of cerebral palsy and the family of function classification systems. ORCID

Kristine Stadskleiv

Researcher at the University of Oslo and Oslo University Hospital, whose work addresses cognitive functioning in cerebral palsy and the valid assessment of cognition under motor constraint. ORCID

Frequently Asked Questions

Is cerebral palsy a progressive condition?

No. The brain injury that causes cerebral palsy is non-progressive, meaning it does not worsen over time. Function can still change with development. It sometimes improves as skills are learned, and sometimes declines in adolescence as growth and secondary problems such as spasticity and contracture take a toll, but the underlying lesion is fixed.

Does cerebral palsy affect intelligence?

It varies. Some people with cerebral palsy have intellectual disability, while many have general intelligence in the typical range. Even with preserved overall intelligence, specific difficulties in attention, working memory, and processing speed are common, and motor impairment can cause cognition to be underestimated by standard tests.

What causes cerebral palsy?

Most cases arise from prenatal and developmental factors, including genetic variants, rather than from oxygen deprivation at birth as was once believed. Prematurity is the largest single risk factor, and the prevalence rises steeply as gestational age falls.

What do GMFCS, MACS, and CFCS mean?

They are three separate five-level systems that grade gross motor function (GMFCS), manual ability (MACS), and communication (CFCS). Because these abilities can dissociate, a person is described by a profile across all three rather than by a single severity score.

How common is cerebral palsy?

It is the most common physical disability of childhood, with a pooled prevalence of about 2.1 per 1,000 live births in population studies. The rate is far higher among children born very preterm or with very low birth weight.

Why is cognition hard to measure in cerebral palsy?

Most cognitive tests require a spoken answer, a pointed response, or a manipulated object within a time limit, which are precisely the abilities cerebral palsy can impair. Unless the test is adapted, a child who knows an answer but cannot produce it can be wrongly scored as having low cognitive ability.

Can cerebral palsy be cured?

There is no cure, because the brain injury is permanent. Management focuses on maximising function and participation and preventing secondary complications, and evidence-graded interventions can meaningfully improve outcomes, especially when begun early.

Why does early diagnosis matter?

The developing brain is most plastic early in life, so intervention begun in infancy may achieve more than the same intervention started years later. Combining neuroimaging with standardised movement and neurological assessment now allows cerebral palsy, or high risk of it, to be identified well before the traditional age.

Support Organizations

References