Background: Very preterm babies are at high risk for developmental disorders. Prechtl's General Movements (GMs) Assessment is a valuable and reliable tool in infants until three months corrected age for predicting their further developmental difficulties, particularly cerebral palsy. However, a specific therapeutic intervention based on this assessment, has not yet been defined. Soloveichick et al. (2019) described the Movement Imitation Therapy for Preterm Babies (MIT-PB) as a promising novel approach based on the model of GMs and the current knowledge of brain development.

Study design and objectives: The present pilot prospective controlled intervention study aims to clarify whether preterm infants born with a gestational age (GA) <32 0/7 weeks showing abnormal GMs at 33-34 weeks postmenstrual age (PMA) differ in their neurodevelopmental outcome at three and 24 months corrected age depending on whether they were treated with usual care physiotherapy or additionally with MIT-PB.

Methods: The participants are recruited at two study sights over 18 months, whereby 40 participants per group are estimated. MIT-PB starts at the NICU and is continued until 52 weeks PMA. The parents are introduced in the method in order to take over a part of the treatment right from the beginning and continue after discharge. The essential content of MIT-PB is to manually guide the infant's abnormal movements into movements as similar as possible to normal GMs.

The primary outcome is the Motor Optimality Score Revised (MOS-R) at three months corrected age. The analysis for the primary endpoint (MOS-R at T2) will be conducted using an analysis of covariance (ANCOVA) with the treatment group (intervention/control) as fixed factor and GMOS at T0 (baseline) as covariate. Several covariates are included in the analysis. Furthermore, neuromotor outcome at term (GMOS) and at two years corrected age (Bayley Scales of Infant and Toddler Development III), as well as parental self-efficacy (Perceived Maternal Parental Self-efficacy tool(PMP S-E)) and dose-response of MIT-PB are evaluated.

1. Background General Movements (GMs) are spontaneous movements that occur in every child from the 9th week of pregnancy until the onset of voluntary motor skills, at the age of five months (Einspieler et al. 2008).

   Normal GMs are characterised as variable, complex, involuntary movements in the neck, trunk and extremities of the fetus and the infant, respectively, variable in amplitude, speed and intensity. According to their age-specific gestalt, GMs are assigned in preterm GMs in preterm born infants, writhing GMs at term equivalent age and Fidgety Movements (FMs) at the age of 52 to 56 weeks' postmenstrual age. As stated by the current literature, GMs allow a good, early prediction of the infant's neuromotor development. Einspieler & Prechtl (2005) report a general sensitivity of 94% with regard to cerebral palsy (CP) or developmental delay at the age of two years. Children with abnormal FMs at the age of three months corrected age, showed delays and abnormalities in the development of motor skills, cognition and language in the Bayley Scales of Infant and Toddler Development III (Bayley III) (Albers et al. 2006). (Peyton et al. 2017) It is supposed that GMs are generated by a Central Pattern Generator (CPG) in the brain stem. The inherently rhythmic activity of the CPG achieves its visible variability through inputs from other parts of the central nervous system (CNS), projections from the cerebellum and basal ganglia, and through feedback from the sensory system and is thereby modulated. The resulting movements (premature and writhing GM) are thought to be primarily involved in the exploration of the movement repertoire and thus sculpt the nervous system (proprioception). Through those CPG-generated movements, all possible combinations of degrees of freedom in the various body joints are explored in order to generate an abundance of variable, self-generated, afferent information (Hadders-Algra, 2018). When a lesion occurs in the CNS, the modulation within the CPG is reduced. The resulting movements are therefore less variable and less complex in their quality, since they reflect the only slightly modulated CPG activations. These movements are referred to as abnormal GMs (Einspieler & Marschik, 2012) Due to the monotonous movements, the sensory afferents, which are necessary for brain development, are less variable as well.

   To date, no specific therapeutic early intervention for children with abnormal GMs starting during preterm age is defined (Anderson et al. 2020; Morgan et al. 2016; Spittle et al. 2012). Current research has shown that key elements in the therapy for preterm babies are the involvement and support of the children's parents to improve the parent-child interaction, a high therapy intensity and the offer of an enriched environment for the child (Anderson et al. 2020; Morgan et al. 2016; Hadders-Algra et al. 2017; Khurana et al. 2020). Furthermore the early intervention should start at the neonatal intensive care unit (NICU) and be continued in the outpatient setting and it should be justified by corresponding assessments at the beginning and throughout the course of the treatment (Anderson et al. 2020; Spittle et al. 2012).

   Dr. Marina Soloveichick's (2019) MIT-PB uses the possibility of modulating the CPG via the sensory afferents in infants with less variability and reduced complexity of movements. The essential content of the treatment method is to manually guide the infant's stiff and cramped movements into movements as similar as possible to normal, fluent, variable GMs. Due to the resulting variability and complexity of the sensory feedback, a modulation of the CPG as normalised as possible, should be influenced via this approach. In the report by Soloveichick et al. (2019), MIT-PB was used in four infants with very abnormal GMs (cramped synchronised). Although they were at high-risk for developing CP, all participants showed normal neurodevelopmental outcome by the corrected age of two, which is consequently attributed to MIT-PB.

2. Hypothesis and specific aims The present study aims to clarify whether preterm infants (born below 32 weeks' gestational age (GA)) showing significantly abnormal GMs at 33-34 weeks' postmenstrual age (PMA) (T0) differ in their neurodevelopmental outcome at three months corrected age (T2) (assessed with Motor Optimality Score-Revised (MOS-R) (Einspieler et al. 2019)) depending on whether they were treated with physiotherapeutic usual care (control group) or additionally with MIT-PB (intervention group). In addition to the neurodevelopmental outcome at T2, it will be examined whether MIT-PB has an influence on the neuromotor outcome at term age (T1) assessed with the General Movement Optimality Score (GMOS) (Einspieler et al. 2016) and 24 months corrected age (T3) (assessed with Bayley III (Albers et al. 2006). Furthermore, it will be examined whether the outcome at 24 months corrected age is related to the total score of GMOS at T0. Finally, the difference in parental self-efficacy from T0 to T2 in the control group versus the intervention group will be investigated (assessed with Perceived Maternal Parental Self-efficacy tool (PMP S-E) (Barnes et al. 2007)).

   It is presumed that there is a statistical relation between the group affiliation and the MOS (adjusted to GMOS) depending on the GMOS at T0. Participants of the intervention group show a greater improvement from the GMOS (T0) to the MOS-R (T2) compared to the control group. The same goes for Bayley III (T3) as well as for the GMOS (T1). It is assumed that participants of the intervention group show a greater improvement from the GMOS (T0) to the GMOS (T1) and to the Bayley III (T3). Likewise, it is hypothesized that the parents performing MIT-PB go through a greater improvement of parental self-efficacy (from T0 to T2) compared to the control group.

3. Experiments to be performed, protocols, methods This study is a pilot prospective controlled intervention study with a comparison between the intervention and control group depending on the initial GMOS at T0.

Intervention Patients of both groups receive usual care physiotherapy during hospitalization, whereby the intervention group receives MIT-PB in addition. Parents are taught how to perform MIT-PB as soon as possible in order to be able to take over a part of the treatment in the inpatient phase and the whole MIT-PB treatment after discharge. The goal is to perform MIT-PB five times a day during ten minutes until the age of 52 weeks PMA. In the outpatient setting, a specialized physiotherapist supports parents from the intervention group. If indicated, patients from both groups receive usual care physiotherapy in the outpatient setting.

Description of the planned statistical methods:

The GMOS and MOS-R assessment scores are not strictly scaled to intervals, but since it is a cumulative score of multiple items, it is assumed that this is the case. The analysis for the primary endpoint (MOS-R at T2) will be conducted using an analysis of covariance (ANCOVA) with the treatment group (intervention/control) as fixed factor and GMOS at T0 (baseline) as covariate.

Sociodemographic and neonatal (major morbidities) are included in the calculation. The secondary endpoints are analysed comparable to the primary outcome analysis with an ANCOVA. As a sensitivity analysis, a dose-response-analysis (ANCOVA) will be conducted for the subgroup in Zurich. Other subgroup analysis will be conducted, if the according analysis shows an interaction.

Inclusion Criteria:

• Preterm infants
• Gestational age < 32 weeks
• GMOS ≤ 23/42 at 33 0/7 - 34 6/7 weeks PMA
• General medical condition stable to the extent that video imaging of the GMs in supine position is possible (minimal underpinning and non-invasive ventilation or oxygen support are allowed)
• Hospitalisation at the NICU at University Hospital Zurich (USZ) or Inselspital Bern
• Place of residence located in order to ensure appropriate follow-up outpatient physiotherapy treatment, if indicated
• Parent or legal guardian must be able to give informed consent as documented by signature

Exclusion Criteria:

• Syndromal disease
• Neuromuscular disease (e.g. myelomeningocele, spinal muscular atrophy)
• Underlying musculoskeletal disease (e.g. arthrogryposis, clubfoot)
• Participation in other studies if it is anticipated that the interventions in each case could affect the outcome of the other study
• Inability of the legal guardian to follow the procedures of the study, e.g. due to language problems, psychological disorders, dementia, etc.

The GMOS and MOS-R, as well as the Bayley III assessors are blinded to the intervention and the group allocation of the participants. The physiotherapists who perform the intervention in both groups do not score the assessments.The statistical analysis is done by a study statistician, who is not part of the core study team.