Inflexible neurobiological signatures precede atypical development in infants at high risk for autism

Abstract

Variability in neurobiological signatures is ubiquitous in early life but the link to adverse developmental milestones in humans is unknown. We examined how levels of signal and noise in movement signatures during the 1st year of life constrain early development in 71 healthy typically developing infants, either at High or Low familial Risk (HR or LR, respectively) for developing Autism Spectrum Disorders (ASD). Delays in early learning developmental trajectories in HR infants (validated in an analysis of 1,445 infants from representative infant-sibling studies) were predicted by worse stochastic patterns in their spontaneous head movements as early as 1-2 months after birth, relative to HR infants who showed more rapid developmental progress, as well as relative to all LR infants. While LR 1-2 mo-old infants' movements were significantly different during a language listening task compared to during sleep, HR infants' movements were more similar during both conditions, a striking lack of diversity that reveals context-inflexible experience of ambient information. Contrary to expectation, it is not the level of variability per se that is particularly detrimental in early life. Rather, inflexible sensorimotor systems and/or atypical transition between behavioral states may interfere with the establishment of capacity to extract structure and important cues from sensory input at birth, preceding and contributing to an atypical brain developmental trajectory in toddlerhood.

Figure 1

Significantly increased subtle fluctuations in spontaneous head movements as timeseries for High Risk (HR) compared to Low Risk (LR) infants, for angular and linear speed. (a) comprises a total of N = 93 datasets (N = 49_HR and N = 44_LR) across all time points, including N = 22 infants tested longitudinally, (b) shows data for 1–2 mo-old infants (N = 28_HR and N = 28_LR), and (c) shows data for 9–10 mo-old infants (N = 21_HR and N = 16_LR). The right-most panels show empirical Cumulative Distribution Functions (eCDFs). The eCDF of the HR distribution was significantly different from the LR eCDF across the entire sample (angular speed: Kolmogorov-Smirnov, p = 1.7935e-28; linear speed: K-S, p = 6.5195e-18) as well as separate for each age group (for 1–2 mo-olds, angular speed: K-S, p = 1.0415e-10; linear speed: K-S, p = 2.4635e-05 and for 9–10 mo-olds, angular speed: K-S, p = 1.3533e-27; linear speed: K-S, p = 3.7688e-15).

Figure 2

(a) HR infants in each age group, 1–2 months (N = 28) and 9–10 months (N = 21), show increased noise-to-signal levels (higher b parameter) and increased randomness (toward more Exponential distributions, lower a shape parameter) relative to age-matched LR infants (1–2 months: N = 28, 9–10 months: N = 16). (b) Considering only 9–10 mo-olds in the longitudinal subset (total N = 22, 11 HR and 11 LR) subgrouped by their noise-to-signal levels as 1–2 mo-olds, LR infants are consistently closer towards more normative ranges on the shape parameter and have lower noise-to-signal levels. Error bars denote 95% CIs.

Figure 3

Linking the rapidity of development with stochastic signatures in infants at a biologically High and Low Risk for developing autism later in life. Parameter estimates from movement data during a resting-state sleep fMRI shown for (a) 9–10 mo-olds: N = 14_HR, N = 13_LR, and (b) 1–2 mo-olds: N = 13_HR, N = 15_LR. Insets show corresponding Gamma probability density functions (PDFs). High Risk (HR) infants with the most delayed (‘stuck’) trajectory show the most deleterious signatures, with their shape parameter tending towards the left on the x-axis (towards more Exponential range), and scale parameter emerging higher on the y-axis (towards increased noise-to-signal or fano factor levels), a relation quantified by power law fits (shown) to the data. For all plots, HR infants who do not make rapid developmental progress show more heavy-tailed, less symmetrical PDFs. Note. Each individual infant’s developmental trajectory is defined independently in a data-driven manner, using exponents of the individual power fits to score data comprising at least 2 or more Early Learning Composite (ELC) scores (for infants with available Mullen scores) collected longitudinally (at 6, 12, and/or 18 months) on the Mullen Early Learning scale. For each risk group, infants are subgrouped into a progress group (those with a higher, more rapidly rising exponent/slope) and a flatter or ‘stuck’ group (those with a lower exponent, indicating a slower rise over time). For Mullen ELC and 9–10 mo-olds data, HRstuck: N = 7_ELC, HRprogress: N = 7_ELC, LRstuck: N = 7_ELC, LRprogress: N = 6_ELC. For Mullen ELC and 1–2 mo-olds data, HRstuck: N = 6_ELC, HRprogress: N = 7_ELC, LRstuck: N = 7_ELC, LRprogress: N = 8_ELC. Error bars denote 95% CIs.

Figure 4

Dissociation of movement signatures to “listen” vs. “sleep” conditions as a function of risk status and age. (a) At 1–2 months, LR infants during sleep show the most symmetric signatures with high signal-to-noise levels yet the noisiest, least symmetric signatures during wakefulness when listening to native language. HR infants show more similar movement signatures during the two conditions. (b) At 9–10 months, HR infants show heightened noise levels during sleep and lower levels during the listening task relative to LR infants. Note. Data points during sleep are the same as those presented in Fig. 2a and included here for ease of comparison relative to those from the listening task. Native Language listening, 1–2 mo-olds: N = 27_HR, N = 28_LR; 9–10 mo-olds: N = 19_HR, N = 14_LR. Sleep, 1–2 mo-olds: N = 28_HR, N = 28_LR; 9–10 mo-olds: N = 21_HR, N = 16_LR. Most of the infants who were scanned during sleep at either time point were also scanned in a separate scanning session while listening to native language speech. Error bars denote 95% CIs.

Figure 5

Biological origins of atypical spontaneous movement patterns during sleep: the role of father’s age. (a) shows 1–2 mo olds and (b) shows 9–10 mo-old infants subgrouped by higher vs. lower Father’s Age (FA) at conception. At both time points, HR infants with an older father (“HR FA High”) have movement signatures characterized by heightened noise-to-signal levels that tend towards Exponential, less symmetric ranges, in contrast to LR infants with a younger father who show the most normative movement signatures. Note. For 1–2 mo-olds with available father’s age data, HR_{FA High}: N = 6, HR_{FA Low}: N = 5, LR_{FA High}: N = 4, LR_{FA Low}: N = 3, and for 9–10 mo-olds, HR_{FA High}: N = 6, HR_{FA Low}: N = 7, LR_{FA High}: N = 3, LR_{FA Low}: N = 2. Error bars denote 95% CIs.