← Cloud analytics

BABA Infant Motor Score (BIMS)

Extensively validated score for quantifying infants' gross motor development

ρ = 0.97 correlation with the Alberta Infant Motor Scale (AIMS)
κ = 0.81 agreement with AIMS when flagging low performers at the 5th and 10th centile
620 unsupervised at-home recordings behind the largest validation, in 134 infants

BIMS is a single number describing how mature an infant's gross motor performance is. It is computed from the second-by-second posture and movement classification of a whole recording, and runs from 0 to 100: near 0, a repertoire of lying supine without yet rolling; near 100, fluent walking as the main way of getting around (Figure 1A).

Because it is not tied to any single milestone, BIMS moves gradually rather than in steps, and it keeps rising after milestone-based scales have levelled off — the AIMS reaches its ceiling several months before BIMS does. That makes it usable as a repeated outcome measure in the same child, read against a normative growth chart (Figure 1B).

AHow a recording becomes a BIMS

BOne infant against the growth chart

Figure 1. A The analysis runs on the recording as a whole: every second is classified, the free-play segments are kept, and the recording's mix of postures and movements is compared with age-specific models from typically developing infants. B Normative BIMS centiles from 580 recordings of 92 typically developing infants, with one infant from the study's external clinical cohort. For the first three recordings the infant's BIMS lay below the 5th centile, while a clinical AIMS assessment also placed them below its 5th centile (filled points). From twelve months the BIMS moved into the typical range, as did the AIMS (open points). Data: Airaksinen et al., Science Translational Medicine 2026, data files S2 (centiles) and S4 (participant id_09).
Show the data for panel B
Age (months)BIMSBIMS z-scoreAIMS at the time
7.05.9−2.56below 5th centile
8.417.3−1.88below 5th centile
10.418.2−3.14below 5th centile
12.256.2−0.78at or above 5th centile
13.983.00.02at or above 5th centile
15.598.80.83at or above 5th centile
18.3100.00.60at or above 5th centile

What you get

Per recording
One BIMS value from 0 to 100, its centile and age-adjusted z-score on the normative growth chart, and the posture and movement distributions it was derived from.
Per child
The individual trajectory across repeated recordings, read against the normative growth chart.
Per study

What you need

Garment
MAIJU suit with a fitting size.
Sensors
Four Movesense Flash or Movesense MD sensors.
Successful recording
From 30 minutes of detected free playtime; the fluctuation settles once a recording passes about an hour.

How it is derived

The classifier output for the whole recording is reduced to posture and posture-conditional movement distributions. Age-specific models of those distributions, estimated from typically developing infants, are each scored against the recording, and BIMS is the likelihood-weighted average of their ages — a developmental age prediction — normalised to the 0 to 100 range.

BIMS correlated with chronological age at r = 0.89 in the original cohort of 59 infants (Communications Medicine 2022), and at ρ = 0.93 across 620 unsupervised at-home recordings from 134 infants, 37 of them in neurological follow-up for an identified or suspected developmental risk (Pediatrics 2025). The normative chart in Figure 1B comes from the growth-chart study, where BIMS was one of 220 motor metrics modelled against age (Science Translational Medicine 2026).

Against the Alberta Infant Motor Scale

A concurrent validity study compared BIMS with the AIMS score given by a paediatric physiotherapist with more than twelve years of experience, assessed within two weeks of the wearable measurement: 67 at-home measurements from 42 infants, 25 of them recruited from a neurodevelopmental follow-up clinic and eight carrying a neurological diagnosis.

BIMS and AIMS correlated at ρ = 0.97. Used to flag low-performing infants at the 5th and 10th centile, the two agreed at κ = 0.81, and a machine-learning prediction of the AIMS score from the wearable data performed no better than BIMS itself (ρ = 0.96). (Developmental Medicine & Child Neurology 2026).

In developmental science

Infant development does not happen in isolation, but exhibits associations between modalities. In a population-based longitudinal sample of 107 infants aged 6 to 21 months, infants whose BIMS ran ahead of their age — measured as an age-adjusted z-score — were also ahead in prelinguistic development (β = 0.033, p = 0.016) and social development (β = 0.038, p = 0.025), both assessed with the Infant Toddler Checklist. Among the individual motor skills, more time in independent movement — crawling, standing and walking — went with more advanced prelinguistic and social abilities (Pediatric Research 2025).

Available on

Publications

All research →

Request a quote