Before self-regulation, there is relationship: what new interoception research asks us to notice from the womb to NICU and perinatal mental health. A new 2026 paper in Scientific Reports has made us reflect on something we often think we understand quite well.
What do we actually mean when we describe a baby as having a “regulation problem”?
Crying, sleeping, and feeding difficulties are familiar in infancy. They can also become separated remarkably quickly in practice. Feeding becomes a nutritional or oral-motor issue. Sleep becomes a sleep question. Persistent crying may be understood through the lenses of action (or behaviour), temperament, attachment, or emotional regulation.
But what if some of these early experiences also tell us something about how a developing body is sensing, predicting and responding to what is happening within and around it?
That is where the new study by Kadiš and colleagues becomes particularly interesting.
Using data from the Bavarian Longitudinal Study, the researchers explored associations between early regulatory problems and outcomes many years later. Regulatory problems were defined as difficulties with crying, sleeping and feeding and were assessed at 5, 20 and 56 months. This allowed the researchers to distinguish between difficulties that were relatively transient and those that were multiple or persisted from infancy into early childhood.
In adulthood, participants underwent MRI and completed measures relating to avoidant personality traits and peer relationships. The final MRI analysis included 47 adults who had experienced multiple and/or persistent regulatory problems and 121 who had not.
The finding that particularly caught our attention was in the insula.
The researchers found greater gyrification, or cortical folding, of the left insula in the group who had experienced early regulatory problems. When they looked more closely, the clearest difference was between people who had experienced persistent regulatory problems and those who had not.
There was another interesting interaction. Left insula gyrification was greatest among people who had experienced early regulatory problems and who also had avoidant personality trait scores above the study’s borderline clinical threshold. Importantly, this was a measure of traits rather than a diagnosis. Other measures of insula structure, including volume, cortical thickness and sulcal depth, did not show the same association. Neither did adult peer relationship quality.
It would be very easy to turn this into an alarming story about unsettled babies and later brain development.
That is not what this research allows us to say.
The study found associations. It cannot tell us that crying, feeding or sleeping difficulties caused later differences in brain structure or psychosocial experience. The authors explicitly acknowledge another possibility. Early regulatory problems may themselves be an early expression of a wider neurodevelopmental trajectory. Genetics, early experiences, relationships and other influences may all be part of a much more complex developmental story.
The authors even raise the possibility that increased gyrification could reflect a different developmental adaptation rather than pathology. That distinction matters.
Perhaps the question this paper leaves us with is not, “How do we make babies better at regulating?” Perhaps it is, “What are we actually observing when a developing body is finding crying, feeding, sleeping or settling difficult?”
The insula brings the body back into the conversation
The insula is particularly interesting because of its role in interoception.
Interoception is sometimes reduced to noticing a heartbeat, hunger or the need to use the toilet. The picture is much richer than that.
The posterior insula contributes to the processing of physiological information arising from within the body. More anterior parts of the insula help integrate internal information with affect, cognition, attention, previous experience and what is happening around us. Kadiš and colleagues describe the insula as part of a broader allostatic interoceptive system that integrates sensory and interoceptive information to support adaptation.
Suddenly, crying, feeding and sleeping begin to look a little different. These are not simply behaviours taking place on the outside of a baby. They involve respiration, temperature, hunger, fullness, fatigue, discomfort, movement, touch, sound, position, pain, familiarity and human contact.
A baby does not consciously divide those experiences into the professional categories we later place around them. They experience them through a developing body. And that body already has a sensory history before birth.
Perhaps the sensory story starts earlier than we sometimes imagine
Before a baby can recognise, name or respond intentionally to hunger, tiredness, discomfort or growing overwhelm, somebody else is already responding.
Research into prenatal auditory experience suggests something even more fascinating. The newborn does not arrive as a sensory blank slate waiting for experience to begin.
In a much-cited 2009 study, Mampe and colleagues compared the cries of 30 French and 30 German newborns. French-exposed newborns more often produced rising melodic cry contours, while German-exposed newborns more often produced falling contours. The authors proposed that the newborn’s cry melody was already reflecting aspects of the speech prosody heard before birth.
This became popularly described as babies “crying in their mother tongue”. It is a wonderfully evocative idea, but newer evidence asks us to hold it carefully.
A 2026 systematic review by van Niekerk, Junge and Chen brought together 18 studies examining cry and other early vocalisations in newborns exposed prenatally to 11 languages. The evidence that newborns consistently reproduce language-specific pitch patterns was sparse and inconsistent. The authors concluded that evidence for prenatal language exposure shaping newborn pitch production remains tentative.
So “babies cry with an accent” is an intriguing possibility, not a settled scientific fact. But the wider story may be even more interesting.
There is stronger evidence that babies perceive and learn from language before birth. Moon, Lagercrantz and Kuhl found differences in newborn responses to familiar and unfamiliar vowel sounds in Swedish and American babies tested, on average, only around 33 hours after birth. Their findings supported the prenatal learning of language characteristics heard in utero.
More recently, René and colleagues used neuroimaging to explore this further. During the final weeks of pregnancy, babies from French-speaking families were repeatedly exposed to a story in French and either German or Hebrew. Within days of birth, their brain responses were measured while listening to these languages. Responses to the foreign language that had been heard before birth more closely resembled responses to French than responses to an unfamiliar language.
The researchers interpreted this as evidence that even relatively brief prenatal linguistic exposure can influence speech processing at birth. That adds something important to our conversation about regulation. A baby arrives in the world already carrying sensory experience.
There has been rhythm, movement, pressure and changing position. There have been maternal body sounds and voice. The melody and cadence of language have reached the baby through the intrauterine environment.
Some experiences have already become more familiar than others. Perhaps, then, before we begin talking about self-regulation, there are other things we need to notice first: familiarity, predictability, sensory history and relationship.
Before self-regulation, there is a relationship
Kadiš and colleagues describe early regulation within a developmental transition from caregiver co-regulation towards increasingly autonomous control.
That sentence is worth sitting with. A baby does not first learn to regulate independently and then discover other people. Another person is already there.
A parent notices a change in movement and adjusts how they hold their baby. Someone recognises a pause in sucking and waits. A voice becomes quieter. Position changes. Light is altered. Touch becomes still rather than moving. Somebody notices that enough is enough before distress escalates.
The baby’s state influences another person’s response, which in turn shapes what happens next. This is why regulation cannot be understood only as an individual ability.
Self-regulation matters. Co-regulation matters. But we have also been using the term “collaborative regulation” to describe the continuously responsive process that occurs between people.
Collaborative regulation (Smith, 2021; Smith et al., 2026) recognises that what becomes possible for one nervous system is continually influenced by the environment, activity, and the nervous systems of other people. It extends the conversation beyond self-regulation and co-regulation by making visible the ongoing, reciprocal nature of regulation within everyday participation.
That is not dependence. It is part of being human. And nowhere is that clearer than in neonatal care.
NICU changes the sensory story
For a baby who arrives early, the expected sensory environment changes dramatically.
Light, sound, gravity, movement, touch, temperature, pain, positioning, hunger, medical procedures, and sleep now occur in neonatal care. At the same time, developmental processes that would have occurred had pregnancy continued were taking place within the sensory environment of the womb.
This does not mean that a neonatal intensive care unit is simply “too sensory”, or that sensory input should automatically be reduced.
That would be far too simplistic.
The more useful questions are individual:
• What sensory information is available, and what is absent?
• What might already be familiar to this baby, and what is unpredictable?
• What are this baby’s earliest signs that they are available for feeding, contact, care or interaction?
• What changes when we alter pace, sound, touch, position or timing?
• How can parents and professionals learn this particular baby’s communication together?
Research examining parental voice in neonatal care offers insight into why this matters.
Filippa and colleagues studied maternal speech during a painful procedure in preterm infants and reported lower pain scores alongside increased oxytocin compared with standard care. In a later randomised trial, maternal speaking and singing were associated with changes in heart rate variability, with the clearest effect during maternal singing. The researchers cautiously suggested that early vocal contact might support the maturation of the autonomic nervous system.
More recently, the 2026 MIND randomised controlled trial combined caregiver voice with scent and supported holding for preterm infants. The researchers reported differences in neural speech sound processing and later language outcomes compared with the control condition.
None of these studies investigated Sensory Ladders®. None demonstrates that a sensory framework changes the development of the insula. That distinction is essential.
But together, these studies make something harder to ignore. Voice is not simply background noise. Touch is not simply handling. Holding is not simply positioning. Sensory experience and relationship are happening simultaneously.
What might a Sensory Ladder® help us notice?
This is where Sensory Ladders® (Smith, 2001) can enter the conversation, not as something this neuroscience has proven, but as a way of helping us notice what the neuroscience makes us curious about.
Sensory Ladders® were developed from clinical practice as a way of making changing sensory and embodied states, early cues and their relationship with participation more visible.
A Sensory Ladder® is not an emotion chart. It is not a behaviour ladder. It is not about deciding that one state is “good” while another is “bad”.
It helps us notice change.
For a newborn, words are not yet available to describe that change, so parents and professionals learn through careful observation alongside clinical monitoring and developmental assessment.
The clues may be tiny.
Breathing changes. Muscle tone shifts. Fingers splay. Sucking pauses. Eyes close. Colour changes. Movement becomes more or less organised. A baby moves either toward or away from contact.
A Sensory Ladder® used within neonatal thinking would therefore not be about placing the baby on a rung and trying to move them elsewhere. It would be about becoming interested in the pattern.
What does this particular baby’s body show when feeding becomes possible? What do we notice before interaction becomes harder? What happens when something around the baby changes? Does the response differ when a familiar person speaks to, touches, or holds them?
And can the parent and professional notice these things together? That last question matters enormously.
The Sensory Ladder® is not something a professional completes about another Person. Its value lies in shared noticing and co-production. With a baby, that necessarily includes recognising and valuing parents’ knowledge of their own child.
There is another body in the room
A parent beside an incubator is also having a sensory and interoceptive experience. They may be recovering physically from pregnancy and birth while trying to understand monitors, alarms, medical language and uncertainty they never expected to encounter.
A monitor sounds, and their heart races before they have consciously worked out why. They realise later that they have not eaten or drunk. Sleep becomes fragmented. Their body remains alert after leaving the unit. They may feel compelled to keep watching. Touch may feel different. Or perhaps they feel unexpectedly numb and struggle to recognise what their body needs at all.
This is where NICU and perinatal mental health cannot really be separated.
If we think only about the baby’s regulation, we risk missing the parent’s embodied experience. If we focus only on the parent’s thoughts and emotions, we can equally miss what their body, senses, environment, and occupations contribute.
A Sensory Ladder® may give some parents another route into language for that experience.
Someone may not initially be able to say, “I am becoming anxious”, but they may recognise, “Everything suddenly sounds too loud”, “My chest changes, “I stop finding my words”, “I don’t notice that I am hungry” or “I suddenly need everyone to move away”.
Those observations are not diagnoses. They are information. They may reveal patterns and offer opportunities to make meaning together.
What happens earlier? What follows? What makes things easier? What increases the demand? What does somebody else do that helps, or sometimes makes things harder?
That feels very different from telling someone to “use their regulation strategies”. It brings the relationship back into regulation.
PEAR TREE™ helps us widen the question
This is also where the PEAR TREE™ Lens (Smith, 2025) becomes useful.
Again, not because the Kadiš study, prenatal language research or NICU studies prove PEAR TREE™.
They do not. The research leads first. Then we can ask whether a framework developed through practice, research thinking and co-production helps us remain curious about what the findings might mean in everyday life.
The PEAR TREE™ Lens was being developed through Smith’s PhD work and had already been informing clinical, family and postgraduate teaching before its formal public launch at the Occupational Therapy Show in November 2025.
PEAR asks us to resist placing the whole explanation inside the individual. For the baby, Person invites us to consider the developing nervous system, medical history, sensory capacities, bodily state and what matters for this particular baby.
The environment draws our attention to sound, light, smell, movement, temperature, equipment, predictability and familiar or unfamiliar people.
The activity asks what the baby is actually trying to do or needs to do. Feeding, sleeping, being held, receiving care, interacting, and simply resting are all occupations with sensory and relational demands.
A relational response then makes explicit what might otherwise disappear from our reasoning.
What does somebody else do next?
Do we recognise early signs of communication, or wait until distress becomes unmistakable? Can we pause? Do we change our voice, touch, timing or position? How much agency can a tiny baby have within care that sometimes has to happen? And how do we support parents to become partners in understanding those responses?
TREE translates that understanding into action through Triage, Relational Intention and Infrastructure, Embodied Evaluation, and Ethics.
The question becomes more than “Did the strategy work?”
We can ask what changed, for whom, at what cost, whether the agreed support could actually be delivered, whether agency and dignity were protected, and whether meaningful participation became more possible.
The same lens can be turned towards the parent.
Instead of stopping at “this parent is anxious”, “hypervigilant”, “withdrawn” or “overwhelmed”, we can remain curious about the Person, their Environment, what they are trying to do, and the Relational response they receive.
- What changes when they are believed?
- What happens when information is slowed down or made clearer?
- What happens when somebody explains the alarm they have just heard?
- What happens when they are supported to eat, sleep or leave the bedside without feeling they have abandoned their baby?
- What happens when a professional notices an embodied change before expecting the parent to translate it neatly into emotional language?
These are not softer questions. They are more complete questions.
So what might this change in practice?
Perhaps the first change is not another intervention.
Perhaps it is what we notice.
For neonatal, early years and perinatal mental health practice, this emerging evidence gives us good reason to remain curious about the relationships between interoception, sensory experience, physiology, familiarity, environment and human response.
It may encourage us to notice early changes in availability rather than waiting for overt distress. It may encourage us to recognise that the sensory environment includes familiar voices and people, rather than focusing only on reducing light and noise. It may remind us to ask about a parent’s bodily experience as well as their thoughts. And it may challenge the assumption that successful regulation should ultimately look like an individual managing everything alone.
There are important limitations.
The Kadiš study is relatively small and cannot establish causation. Its sample was entirely white Caucasian, limiting how confidently the findings can be generalised.
The story of babies “crying in their mother tongue” remains intriguing rather than settled. The latest systematic review found inconsistencies between studies and concluded that evidence about language-specific newborn pitch production remains tentative.
The neonatal studies also have their own methodological and sample limitations. None of this evidence allows us to say that Sensory Ladders® or the PEAR TREE™ Lens alter neurodevelopment. We should not ask the research to prove something it did not investigate. But perhaps research does not always need to give us a programme. Sometimes its most useful contribution is a better question.
Perhaps regulation is not quite what we thought
We often ask, “How can we help this person regulate?” This research invites us to widen that question.
What is this body experiencing? What has it experienced before? What is familiar and what is unpredictable? What is happening around the person? What are they trying to do? What happens when another human responds? And what could we change together so that participation becomes more possible?
For a baby in the NICU, participation may mean being available for feeding, sleep, touch, interaction or the comfort of a familiar voice.
For a parent, participation may mean being able to hold their baby, eat, rest, make a decision, talk with the medical team, tolerate the journey into the unit, ask for help, leave and return.
Those occupations matter.
The new interoception research does not tell us to become better at identifying “poor regulators”.
Perhaps it asks almost the opposite.
To notice bodies without reducing people to them. To recognise persistent difficulty without turning it into a prediction. To listen to communication before deciding what behaviour means. To recognise that sensory learning and familiarity begin before birth.
And to remember that while self-regulation develops, it develops within experience, environment and relationship. Participation begins in the body.
But the body was never alone.
References
- Kadiš, M., Wolke, D., Tsalacopoulos, N., Sorg, C. and Mulej Bratec, S. (2026). Associations between early regulatory problems and long term neurobiological and psychosocial outcomes. Scientific Reports. DOI:
https://doi.org/10.1038/s41598-026-66979-z Article: https://www.nature.com/articles/s41598-026-66979-z - Mampe, B., Friederici, A.D., Christophe, A. and Wermke, K. (2009). Newborns’ cry melody is shaped by their native language. Current Biology, 19, 1994–1997. DOI: https://doi.org/10.1016/j.cub.2009.09.064 PubMed: https://pubmed.ncbi.nlm.nih.gov/19896378/
- Moon, C., Lagercrantz, H. and Kuhl, P.K. (2013). Language experienced in utero affects vowel perception after birth: a two country study. Acta Paediatrica, 102(2), 156–160. DOI: https://doi.org/10.1111/apa.12098 Article: https://onlinelibrary.wiley.com/doi/10.1111/apa.12098
- René, A., Caron Desrochers, L., Tremblay, J. et al. (2025). Prenatal linguistic exposure shapes language brain responses at birth. Communications Biology, 8, 1155. DOI: https://doi.org/10.1038/s42003-025-08594-8 Article: https://www.nature.com/articles/s42003-025-08594-8
- van Niekerk, E., Junge, C. and Chen, A. (2026). The influence of prenatal language exposure on the use of pitch in newborns’ vocalisations: a systematic review. First Language, 46(2). DOI:
https://doi.org/10.1177/01427237251372203 Article: https://journals.sagepub.com/doi/10.1177/01427237251372203 - Filippa, M., Monaci, M.G., Spagnuolo, C. et al. (2021). Maternal speech decreases pain scores and increases oxytocin levels in preterm infants during painful procedures. Scientific Reports, 11, 17301. DOI:https://doi.org/10.1038/s41598-021-96840-4 Article: https://www.nature.com/articles/s41598-021-96840-
- Filippa, M., Nardelli, M., Sansavini, A. et al. (2024). Maternal singing sustains preterm hospitalized newborns’ autonomic nervous system maturation: an RCT. Pediatric Research, 95, 1110–1116. DOI:
https://doi.org/10.1038/s41390-023-02932-4 Article: https://www.nature.com/articles/s41390-023-02932-4 - Maitre, N.L., Kjeldsen, C.P., Jeanvoine, A., Lukemire, J., Slaughter, J.L. and Key, A.P. (2026). The MIND Randomized Controlled Trial: An Intervention to Improve Neural Speech Processing and 2 Year Language Outcomes of Infants Born Preterm. The Journal of Pediatrics, 296, 115187. DOI: https://doi.org/10.1016/j.jpeds.2026.115187 Free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC13365924/
- Smith, K. (2001). Sensory Ladders®: Regulation, Perception, Praxis and Participation. The Sensory Project. Developed through clinical practice and co-production in Cornwall. Sensory Ladders®: https://sensoryproject.org/sensory-ladders/ Sensory Ladders® history and resources: https://sensoryproject.org/sensory-ladders/news-sensory-ladders/
- Smith, K. (2025). PEAR TREE™ Lens: A shared participation framework for practice. ASI Wise and The Sensory Project. Developed in earlier clinical and family practice and within workshops and postgraduate teaching, and formally launched at the Occupational Therapy Show in November 2025. A Pocket Guide including PEAR TREE™ Lens: https://sensoryproject.org/pear-tree-model/