'Processing of prosodic phrase boundaries by preterm newborns'
(AsPredicted #198015)


Author(s)
This pre-registration is currently anonymous to enable blind peer-review.
It has 5 authors.
Pre-registered on
11/07/2024 01:02 AM (PT)

1) Have any data been collected for this study already?
It's complicated. We have already collected some data but explain in Question 8 why readers may consider this a valid pre-registration nevertheless.

2) What's the main question being asked or hypothesis being tested in this study?
Research question: do preterm born infants rely on innate mechanisms when processing major prosodic boundaries at birth?
Hypothesis: at birth, preterm born infants can already process prosodic boundaries, which suggests they use innate perceptual biases in the form of the Respiratory Code (RC) and the Iambic Trochaic Law (ITL). According to the Respiratory Code, high pitch is associated with phrase beginnings and low pitch with phrase endings, and silent intervals occur between phrases. According to the ITL, lengthened elements occur phrase-finally. Based on these biologically motivated principles, preterm newborns are able to process prosodic phrase boundaries through pitch change, final lengthening and pauses as acoustic boundary cues.
[Specific predictions based on this hypothesis are provided in section 8]

3) Describe the key dependent variable(s) specifying how they will be measured.
The key dependent variable is mean amplitude of the Closure Positive Shift (CPS) event-related potential (ERP) component in µV. The CPS is an electrophysiological response to prosodic boundary processing during speech perception, that can be measured using electroencephalography (EEG). This response occurs between 400 and 800 ms after onset of a pre-boundary syllable in adults and six-month-old infants. The polarity and region of interest may be different in preterm infants, and may depend on the chosen reference. The mean ERP amplitude will be measured for each of the conditions described in section 4, using BrainVision Analyzer Software. EEG will be recorded using the eego EEG acquisition system (ANT Neuro).

4) How many and which conditions will participants be assigned to?
Each infant is presented with speech stimuli, while EEG is being recorded. The stimuli consist of acoustically manipulated tokens of a Dutch utterance. The utterance consists of a sequence of three coordinated names ("Moni en Lilli en Manu"). These stimuli are presented in a pseudo-random order in 6 conditions involving two readings of the sequence, i.e. [Moni en Lilli en Manu] vs. [Moni en Lilli] [en Manu]. These two readings are achieved by the absence or presence of a major prosodic boundary, i.e. an Intonational Phrase boundary, after the second name in the name sequence. Perception of this boundary can be triggered by specific boundary cues. The six stimulus conditions differ in prosodic boundary cues on the final syllable of the second name in the sequence ("-li"):
1. No cues (control condition)
2. All cues (pitch rise, final lengthening and pause)
3. Pause (following the final syllable of the second name)
4. Pitch fall
5. Pitch rise
6. Final lengthening

5) Specify exactly which analyses you will conduct to examine the main question/hypothesis.
We will use linear mixed-effects modelling to assess whether infants are able to process prosodic boundaries marked by different prosodic cues. ERP amplitude will be used as outcome variable, participant as random factor, and condition (1-6), postconceptional age, and the interaction term age x condition as fixed factors. In order to examine whether the CPS response was obtained in each individual condition, we will compare each boundary condition (2-6) to the no boundary condition (1).

6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations.
A baseline correction of 200 ms prior to the timelock will be applied to the calculated ERP. Next, channels containing artefacts will be removed from each trial using the following rejection criterion: channels exceeding an absolute maximal amplitude of 50 µV during the entire trial and/or a maximal amplitude difference of 30µV per 20 ms and/or a minimal amplitude of 0.5 µV during a 100ms interval are rejected. Channels with at least 10 trials remaining after artefact rejection are subsequently re-referenced to a common average, which includes all channels that meet the 10-trial minimum. If <50% of all electrodes do not meet the 10-trial criterion in any condition, that participant is excluded from further analysis. Finally, average ERPs will be calculated over the remaining data on an individual channel basis.

7) How many observations will be collected or what will determine sample size?
No need to justify decision, but be precise about exactly how the number will be determined.

We aim to recruit 50 preterm infants born between 28 and 33 weeks of gestation with a monolingual Dutch home environment. We will allow for transparent intermediate data review to determine whether to continue or halt participant recruitment before these numbers are reached. Specifically, when data collection for half of the intended sample size is complete, a preliminary EEG analysis will be carried out to check both the validity of the preprocessing pipeline and the presence of the N1-P2 complex in the ERP data, as a proxy for low-level acoustic stimulus perception. Regarding observations per participant, each infant is presented with 60 acoustic stimuli per condition, i.e. 6 x 60 = 360 trials per included infant.

8) Anything else you would like to pre-register?
(e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?)

• Some data has already been collected, but no formal analyses have been performed yet.
• We predict that, within one week after birth, infants born between 28 and 33 weeks of gestation are able to process prosodic phrase boundaries on the basis of the RC and the ITL when perceiving prosodic boundary cues. According to the RC, higher-pitched elements occur phrase-initially, lower-pitched elements occur phrase-finally, and silent intervals occur between phrases. According to the ITL, shortened elements occur phrase-initially and lengthened elements occur phrase-finally. This leads to the following predictions:
- Preterm born infants are able to process intonational phrase boundaries that are marked by a pitch fall, final lengthening and pauses as single boundary cues and prosodic boundaries that are marked by all three cues.
- Preterm born infants are not able to process intonational phrase boundaries that are marked by pitch rise or no cues.
• In order to demonstrate that the mechanisms involved in processing prosodic phrase boundaries are innate rather than dependent on linguistic experience, we will test preterm infants in a broad age span from 28 to 33 weeks
• Preterm newborns' brains are developing significantly between 28 and 33 weeks of gestation. At 28 weeks, the newborn brain may not yet be matured sufficiently in order to generate stable and replicable CPS responses. However, at 33 weeks, the newborns' prosodic processing abilities may already have been influenced by their prenatal linguistic experience. Effects of gestational age on CPS amplitude in different conditions will therefore be analyzed by including the newborns' postconceptional age as a continuous variable in our statistical model.
• Following positive evidence from Dutch-exposed preterms, we will replicate findings in preterms born between 28 and 33 weeks of gestation with a monolingual German home environment.
• It may the case that the ERP effects are delayed and cannot be observed in the usual time window for the CPS component, because brain myelination is still in a very early stage in preterm newborns. Therefore, the time window for analysis will be determined based on visual inspection of the ERP.

Version of AsPredicted Questions: 2.00