CP isn’t just one diagnosis: clinicians use standardized systems to capture motor type, severity, manual ability, and communication function. The three main tools (GMFCS, MACS, CFCS) turn a broad diagnosis into a specific, actionable profile.
Four systems do the work, each with five levels. GMFCS grades gross motor function and is reliable from age two. MACS covers how a child handles objects and is validated from four to eighteen. CFCS covers everyday communication, and EDACS covers eating and drinking. A diagnosis names the condition. These name what a person actually does.
That is the whole reason they exist, and it is why a clinic letter carrying a level is worth more than one carrying an adjective.
Four separate scales are in routine use, each answering a different question: how a person moves, how they handle objects, how they communicate, and how they eat and drink. They are deliberately independent, because those abilities do not track together. A child who walks unaided can have a hand they barely use. A child transported in a wheelchair can have unimpaired language.
What follows is what each scale measures, how well it agrees between raters, and where the older topographic vocabulary still sits. The parent guide is the five types of cerebral palsy.
The best way for parents to get a clear picture of their child’s limitations is to request evaluation using all four of the current classification scales.
Three motor types are formally recognized: spastic, dyskinetic and ataxic. Everything else in common use is either a description or a leftover.
Spastic cerebral palsy accounts for around 80% of cases according to the CDC, and its defining feature is velocity-dependent tone: a limb moved slowly can feel close to normal and the same limb moved quickly catches. Dyskinetic CP produces involuntary movement that increases with effort and emotion and disappears in sleep. Ataxic CP, the rarest, comes from cerebellar involvement and shows as a wide unsteady gait and a tremor that worsens as a hand approaches its target.
Mixed is a description used when a child’s movement crosses those boundaries, not a fourth category, and registers ask for a dominant type to be assigned anyway. A hypotonic type appears on some lists; low tone is common in infancy and usually resolves into one of the three, so it reads better as a stage than a destination. The full picture is in spastic cerebral palsy and its siblings.
Diplegia, hemiplegia, quadriplegia. The familiar names are the least reliable part of this whole apparatus, and there is a study that measures exactly how unreliable.
Thirty clinicians from the Surveillance of Cerebral Palsy in Europe network classified the same ten written case vignettes. Agreement on whether a case counted as cerebral palsy at all was moderate, with a kappa of 0.59. Assigning subtype was worse, and the two areas they most often diverged on were spastic against dyskinetic, and how to describe the distribution of spastic involvement. Twenty of them repeated the exercise five months later: inclusion held up at 0.72, while subtype and GMFCS assignment varied considerably.
European registers responded by collapsing the topographic names into two. Spastic cerebral palsy is recorded as unilateral, which covers hemiplegic CP and a single affected limb, or bilateral, which covers diplegia and quadriplegia. Two categories clinicians agree on carry more information than five they argue about, even though the five sound more precise.
Tone itself is graded on the Modified Ashworth Scale, six grades running 0, 1, 1+, 2, 3 and 4, where 0 is no increase and 4 means the limb is rigid. It measures resistance to passive movement, which is a different question from what a person can do, and it is worth keeping the two apart when reading a report.
These two are classified by what the movement looks like rather than by which limbs are involved, which is why the topographic vocabulary was never a good fit for them.
Dyskinetic cerebral palsy divides into dystonic, where sustained contractions twist a limb into an abnormal posture, and choreo-athetotic, where movements are writhing or irregular and unpredictable. The distinguishing features are behavioral rather than anatomical: the movements increase with voluntary effort and with emotion, and they disappear during sleep, which means a child performs worst at exactly the moment they are trying hardest.
Ataxic cerebral palsy is the smallest group and the one most often reclassified, because low tone and poor coordination in infancy can resolve into a different pattern or turn out to be a different diagnosis altogether. Progressive cerebellar conditions have to be excluded before ataxia is attributed to cerebral palsy, since cerebral palsy is non-progressive by definition and a worsening course points elsewhere.
Diagnosis and classification are separate jobs on separate timetables. The diagnosis can be made in the first months of life. The classification waits until there is enough behavior to classify.
The 2017 international clinical practice guideline in JAMA Pediatrics established what predicts cerebral palsy accurately in infancy: term-age MRI at 86% to 89% sensitivity, the Prechtl Qualitative Assessment of General Movements at 98%, and the Hammersmith Infant Neurological Examination at 90%, all before five months corrected age. The same guideline records that topography and severity are harder to determine that early, which is the reason a subtype written at ten months often is not the one written at four years. See how cerebral palsy is diagnosed and the role of MRI.
Each answers a different question, and each has five levels:
The Gross Motor Function Classification System runs from level I, walking without limitation, to level V, transported in a manual wheelchair with limited head and trunk control, and it is valid and reliable from two years of age. The Manual Ability Classification System does the equivalent job for how a child handles objects in daily life and is validated for ages 4 to 18. It matters most where the two disagree: a child with one affected arm usually sits at GMFCS level I, a rating that says nothing at all about the hand.
The Communication Function Classification System covers everyday communication in both directions, sending and receiving, and it counts any method including a device. The Eating and Drinking Ability Classification System covers eating and drinking, which is worth classifying formally because safety rather than skill is what is being described. Where those levels sit shapes physical therapy, occupational therapy and the case for surgery more than the diagnosis does.
Reasonably, and not perfectly, and a page recommending them should say so. When the CFCS was validated, interrater reliability was 0.66 between two professionals and 0.49 between a parent and a professional, improving to 0.77 for children over four. That is the kind of number worth knowing before a level is treated as a fixed fact about a child, particularly when a parent and a clinician disagree, which the data suggests will happen fairly often.
The practical response is not to discard the scales. It is to ask who classified, when, and on what basis, and to expect a level to be revisited as a child grows. Levels also carry real predictive weight once assigned: severity levels covers the motor development curves and the hip displacement risk that follow from a GMFCS level.
Four function scales in routine use, each with five levels and each answering a different question: GMFCS for gross motor function, MACS for handling objects, CFCS for everyday communication and EDACS for eating and drinking. Alongside them sits the older topographic vocabulary, which European registers have collapsed into unilateral and bilateral.
It sorts children by what they do rather than what they cannot, from level I, walking without limitation, to level V, transported in a manual wheelchair with limited head and trunk control. It is valid and reliable from two years of age, and a level carries real predictive weight: five motor development curves were built from 657 children, one per level.
Because walking and hand function do not track together. A child with one affected arm usually sits at GMFCS level I, and that rating says nothing at all about the hand. MACS grades how a child handles objects in daily life across five levels and is validated for ages 4 to 18, so it catches what the motor scale misses.
Whenever communication is part of the plan, which is most of the time. It covers everyday communication in both directions, sending and receiving, and counts any method including a device, so it describes how well a person communicates rather than whether they speak.
They replace adjectives with something that predicts. A level ties to motor development curves, to hip displacement risk running from 0% at GMFCS I to 64% at level V, and to a surveillance schedule set by age and level. Mild, moderate and severe predict nothing and hide the differences that matter.
Reasonably rather than perfectly. When the CFCS was validated, interrater reliability was 0.66 between two professionals and 0.49 between a parent and a professional, improving to 0.77 for children over four. Subtype classification is weaker still: 30 clinicians classifying the same ten vignettes agreed on inclusion at kappa 0.59 and diverged most on the distribution of spastic involvement.
Kelsey is an experienced surgical nurse who works extensively with pediatric patients and families navigating birth injuries and cerebral palsy diagnoses. She is licensed by the Missouri Board of Nursing and operates under the Nurse Licensure Compact, allowing her to practice across compact-member states. As both a mentor and patient advocate, Kelsey focuses on safety, communication, and compassionate care: helping families understand medical procedures, treatment options, and the realities of pediatric conditions.
View Kelsey Pabst, RN's full bio →
Our nurses, patient advocates and legal experts are solely focused on bringing you the latest cerebral palsy information, options for financial assistance and access to community support.
Editorial process →The information on this page is provided for general educational purposes and is not a substitute for professional medical or legal advice, diagnosis, or treatment. Always seek the guidance of a qualified physician or attorney with any questions about a medical condition or legal matter. Read our full disclaimer.
