Emotional availability: detecting human connection and links with healthy functioning in infancy and early childhood

Introduction

Emotional availability (EA) in adult-child relationships is conceptualized as “the connective tissue” in relationships (Biringen et al., 2014). It doesn't matter what the adult or the child does, if it is not affectively received by the other—and hence, the dyadic nature of any human connection. EA dimensions include four qualities on the adult side (sensitivity, structuring, nonintrusiveness, and nonhostility) and two qualities on the child side (responsiveness to the adult and involvement of the adult). While there are adult and child sides to the relationship, each is relational or dyadic. Specifically, each quality is possible only in the context of a dynamic, mutually reinforcing feedback loop, such that it is nearly impossible for a mother (or father) to be highly sensitive to a child who is constantly rebuffing, except perhaps in a therapeutic way. It is nearly impossible for a teacher to structure a child who does not want to be structured and puts on blinders to the eyes and headphones to ears. And, so on, at least for most of the parent scales. However, it is very possible to be intrusive and hostile to a child “out of the blue” and without child provocation. Thus, the dyadic aspect of the relationship may not be as strong for these two scales, at least at low/ negative end. For the child's side of emotional availability, the child cannot respond unless there are adult initiations and the child cannot involve much if the parent is constantly intrusive.

The EA System now also includes the emotional attachment zones (“EA Zone”) for each member of the relationship (Saunders et al., 2015, 2017; Wurster and Biringen, 2020), based predominantly on the adult sensitivity and child responsiveness scales. With that said, scales such as structuring, nonintrusiveness, and nonhostility being extremely low also guide the assignment of the lowest zone (Problematic). Most prominently, a parent being hostile (especially if overtly hostile) is reason to consider the lowest EA zone, that is, Problematic/Traumatized/Traumatizing, given that frightened or frightening behavior is a hallmark of disorganized attachments (Hesse and Main, 2006).

Brief description of the methodological context and training for EA observations

Emotional availability (EA) is evaluated typically in a 15–20 min, videotaped episode. Longer observations are precious, while less than this recommended amount tax a trained EA evaluator's confidence about what is observed and may compromise reliability and/or validity. Though risky, some research indicates that less observation time can still yield some valuable information (cf. Biringen et al., 2014). In clinical work (where serious recommendations will be made to parents or agencies), significantly more time is needed, even with the availability of videotape.

Nondirective instructions are recommended, such as “go about your normal day with your child” or “be with your child as you normally would”, rather than “play with your child”. Highly structured sessions (e.g., a few minutes of teaching, a few minutes of play, and so on) yield less reliable and valid results than sessions that are semi-structured, where the parent and child can be as they normally present. An EA observation setting echoes a Rorschach test, where the participants interpret the instructions in a way that is unique to their own way of being and relating.

EA is most typically used to evaluate interactions from ages 0–17 years, with the greatest research focusing on infancy and early childhood and newer research focusing on school age and adolescence. Thus, the construct and its measurement system is used without change of the observational tool throughout the periods of infancy, early childhood, middle childhood, and adolescence. One manual focuses on 0–5 years, while a second focuses on 6 years and up. This change in manuals is simply to address elaboration that can help coders, but the conceptualization remains unchanged. More recently developed is the Prenatal EA measure (Salo et al., 2019), which can be measured observationally during the last trimester of pregnancy, after the mother feels the baby moving. Only sensitivity and nonhostility are assessed. More research is needed in this area, but thus far, the Prenatal EA observational measure appears to be related to internal working models of attachment (Salo et al., 2019). Whether it is actually related to postnatal EA remains to be assessed. There are five reviews that have thus far been published (Biringen et al., 2014, 2022, 2023; Clark et al., 2021; Lotzin et al., 2015), and even in these reviews, we were unable to include the close to 500 research articles on EA.

For all age groups, each of the six EA Scales are rated on 7-point scales, with 7 being optimal and 1 being inadequate. Another way of evaluating EA allows for rating on longer dimensions (i.e., 1–29 points), which can be useful for intervention or therapeutic work so one can see how much improvement has been happening over time. This is a challenging system to learn and authorized training is required. Even upon certification, the user acquires a minimum level of competence that allows them to go forward with their data coding and checks on inter-rater reliability in their own samples. True mastery of the system requires extensive use and collaboration with other EA users over time.

EA observations are held in either stressful or non-stressful contexts, and most commonly real-world settings (e.g., waiting room, park, playroom). There are no required toys or materials, although most observation settings include some age-appropriate toys or other materials. As noted, the observations are videotaped, mainly for ease of achieving inter-rater reliability. I know of only one study that conducted a brief direct observation (e.g., McCarthy et al., 2003). Most EA observations require that the face of the parent and child be clearly visible at least a good part of the time (though not constantly), since the context can be naturalistic; what they say to one another should ideally be translated. A small number of studies have been able to do coding of EA without translations, as for example when there are 17 languages in a sample and it was not possible to find interpreters (e.g., Lebiger-Vogel et al., 2022). One group had to film in refugee camps in which it was impossible to actually hear what the parents and children were saying to one another (Scharpf et al., 2023). These variations that have yielded meaningful results indicate that EA is a robust and user-friendly tool, but limiting the incoming nonverbal and/or verbal content is likely to create some changes in nuance perception. For example, I do not think that anyone has published results from the covid pandemic where families were actually wearing masks and only the eyes, voice, and body positioning could be used as emotion cues. If investigators and/or clinicians begin to limit duration of time, understanding of verbal content, and most importantly, nonverbal (facial) content, there is significant risk that evaluations may be based on coarse rather than nuanced information.

The six EA scales

As described above, each of the six scales is rated on interval scales. Adult sensitivity refers to the adult's ability to be present in a healthy emotional connection with the child. It is a simple quality, and yet, it can be challenging to evaluate. This is because the influence of behaviorism has been so overwhelming in our field—such work indicates that more nice things you say and the more nice things you do, the better a parent you are (Troutman, 2015). Actually, this is not the case; good words and deeds have to be genuine and authentic (and not necessarily bountiful) to reach and connect in a sensitive way with a child. Lower levels of sensitivity typically refer to detachment or dysregulated behavior. A good deal of video-based training is needed to be able to objectively observe both security/emotional availability as well as insecurity/emotional unavailability.

Adult structuring refers to an appropriate level of guidance and success in reaching the child. Overdoing it (i.e., overstructuring) is not helpful to the child and suggests that the adult is not reading the child's cues. If it does not reach the child, we may actually be seeing overstructuring. From a behavioristic standpoint (Troutman, 2015), the more one does, the better it is, but from an emotional availability standpoint, the structuring has to hit the mark and be dyadically successful. Yes, the parent may be structuring a lot, but only when it reaches the child is it appropriate structuring. At the lower end is unstructuring behaviors and at the extreme it can seem like neglect.

Adult nonintrusiveness refers to an adult who is able to let the child lead and is able to follow that lead (Biringen et al., 2014). True, children need to learn many things and adults know a great deal that can be taught to children. Yet, children have masterful ways in which to resist such presumptuous leading, as for example, passive aggressive rejection of overtures, simple ignoring of the adult, or passive/nonparticipatory acceptance of interferences. Beyond these “benign” exchanges, when an adult becomes physically intrusive into the child's space, then we lower the ratings significantly. However, it is challenging to assess whether a physical move into another's space is intrusive or actually structuring (e.g., a mother helping her baby with spoon feeding or holding the infant's back as he is learning to be steady on his feet). In the EA System, context, age-appropriateness, and whether there is actually structuring going on are important pieces of the nonintrusiveness puzzle.

Adult nonhostility refers to a calm, peaceful presence without the threat of negativity (Biringen et al., 2014). Stressful, frustrated ways of being are described in the EA System as covert hostility. When there is some hostility, but it is just bubbling at the surface, not really expressed, it is covert hostility. On the other hand, when the hostility is clearly expressed (yelling, negative/demeaning statements to the child or others), it is overt hostility. A curious aspect of the EA System is that hostility shown to another in the child's presence (e.g., a phone call) are part of what is referred to as “background anger” and included as hostile acts, because the child is exposed to them (Biringen et al., 2014). Most observations are typically in low-stress situations in which the adult member of the relationship can put one's best foot forward. Even if not explicitly requested, the “demand characteristic” of being filmed suggests being at one's best. Thus, both hostility and intrusiveness are low incidence events in low-stress, low-frustration contexts and so their emergence in an observation session are noteworthy.

Child responsiveness refers to the child's affect and behavior (Biringen et al., 2014). A child who is generally satisfied or even happy would likely be on the high end of emotional availability. Could such a child at times be nice to the parent? Yes, we want our children to be nice to us and such behavior is considered to tip toward solicitousness or parentification if it is embedded in a larger psychopathology context. A child simply complimenting the parent on occasion in an otherwise healthy interaction is what all parents would like to experience and is not a sign of people pleasing or parentification. It is important to keep in mind what Emde and Easterbrooks (1985) referred to as the “affective barometer” of a relationship. As was described for sensitivity, significant video-based training and exemplars are needed for experienced professionals to identify security/emotional availability as well as insecurity/emotional unavailability.

Child involvement of the adult refers to the child's tendency to not just respond (because the other is initiating), but to take the initiative (Biringen et al., 2014). If involving behaviors are needy, anxious, or negatively tinged in some way, then a lower rating is assigned. But if the child is under-involving or un-involving, an even lower rating is assigned.

Emotional attachment zones

The four EA zones (“emotionally available,” “complicated,” “detached,” and “problematic/disturbed,”) theoretically map onto the traditional four attachment categories of secure, insecure/anxious, insecure/avoidant, and disorganized are respectively. The zone placement for the parent is predominantly dependent on the parent's sensitivity scale, and the zone placement for the child is predominantly dependent on the child responsiveness scale. Using three separate samples, the zones were found to be significantly correlated with attachment (as measured by either the Adult Attachment Interview, the Attachment Q-Sort, or the Strange Situation Procedure (Wurster and Biringen, 2020). Interestingly, within each zone, there is dimensionality such that we can describe the extent to which an individual is more or less detached, more or less emotionally available, and so on. Also interesting is that adult and child can be in different insecure/emotionally unavailable zones (e.g., one detached and the other complicated), but it is challenging for one to be emotionally available and the other not, except in special populations, such as adoption, foster care, or interactions between a child and a grandparent or with a nanny. The reason for this is that EA is a dyadic concept and it is challenging to imagine how one can maintain an emotional connection with the other without similar reciprocity.

Imagine a symphony orchestra as an analogy for EA. The conductor is the emotionally available parent and the orchestra is the child side, involving a diversity of cues and expressions. Each musician in the orchestra plays a unique instrument, suggesting the different facial expressions, acoustic sounds, and body language of the child. Just as the maestro orchestrates the resulting ensemble, the parent interprets each nonverbal cue of the child and harmonizes responses to fit the child's needs at that moment. This involves sensitivity to subtle emotional changes, appropriate structuring and leading or following when needed, while playing down sudden shifts and intrusions or becoming stressed during challenging moments. If the conductor (parent) creates cacophony instead of this beautiful harmony, interventions for conductor skills may be needed.

Can EA be self-reported?

My professional opinion based on not only the published data, but also prepublished data (that may or may not be published in the future), the link between observations and reported emotional availability is not consistent or strong. Therefore, I choose not to review this in detail here and refer to Biringen et al. (2023). For most research and clinical endeavors, nothing can take the place of astute and nuanced observations. However, this self-report and other self-report measures have their place—in telling us about the parents' feelings and perceptions, which in many cases, may not match professional opinion about the interaction. Yet, there is a place in combining such observations with reported information (Biringen et al., 2023).

Healthy functioning in infancy and early childhood: relations to EA

Mental health during infancy and early childhood is most easily seen in relational terms, including social development (e.g., child-to-parent attachment, the child's emotional availability toward parents, and the child's emotion regulation). In addition, health during infancy and early childhood is likely to be measured in terms of brain mechanisms, stress physiology and neuroendocrine system, and daily activities such as feeding/eating/maintaining a healthy weight. We take each in turn and then move to EA and brief prevention/intervention programming.

Early social/emotional development

The child's emotion regulation

However, this is likely not be the whole of the explanation as EA and emotion regulation in the immunization context have been reported (Pillai Riddell et al., 2011), but not consistently (Racine et al., 2012), suggesting that this context is about more than mother-child emotion communication. Dyads with higher EA demonstrated greater emotional control in a challenge context (especially maternal nonhostility) (Little and Carter, 2005). EA has been found to be predictive of infant emotion regulation (Martins et al., 2012). Of course, there is the possibility that it is easier to be higher in EA with a child who is appropriately controlled.

Infant-parent attachment

The parental side of EA has consistently been found to be related to attachment security, be it measured by the strange situation (e.g., Aviezer, 2008; Wurster and Biringen, 2020) as or the Attachment Q-Sort (e.g., Altenhofen et al., 2010, 2013), which are the leading tools to date. Furthermore, the magnitude of the correlations is about the same as these attachment measures to one another (Wurster and Biringen, 2020). In some studies, specific aspects of EA (e.g., nonintrusiveness) is linked with disorganized attachment (Swanson et al., 2000), suggesting the importance of a full EA evaluation rather than only the sensitivity scale. The reader is also referred to Biringen et al. (2014) for a fuller description of these studies.

The child's availability to parents

The parental side of EA is closely linked with the child's side, and in particular the child responsiveness scale (Biringen et al., 2014, 2022, 2023). However, this is not always the case in children with disabilities (e.g., autism spectrum disorder, ASD), where a parent's emotional availability is met with less availability from the child (Oppenheim et al., 2025). Yet, the research indicates that in ASD the parent and the child sides are correlated and the child does show a full range of the child's side of emotional availability, especially in structured situations (Oppenheim et al., 2025).

Physical and physiological development

Brain mechanisms

Miller et al. (2022) found that (maternal) EA following premature birth provided an opportunity to promote optimal neurodevelopment in children, as measured by magnetic resonance imaging. Very young infants' (0–6 months) greater neural sensitivity to negative emotions was associated with the mother's intrusiveness (Kim et al., 2017). Copeland et al. (2022) reported that early maternal sensitivity at 8 months predicted children's local brain connectivity at 5 years of age. However, it is unclear why early sensitivity (but not sensitivity in toddlerhood, 30 months) was predictive, suggesting the importance of “early”. Taylor-Colls and Pasco Fearon (2015) examined parental sensitivity and 7 month olds' neural responses to emotional faces, using event-related potentials (ERPs), with maternal sensitivity related to higher amplitude of infant responses to positive facial expressions.

Stress physiology and the neuroendocrine system

Kertes et al. (2009) found that (maternal) EA is associated with lower stress response, as measured by hypothalamic pituitary axis (HPA) stress response in the mornings. Children of immigrants (in Germany) also showed lower HPA axis response at kindergarten entry if their mothers were higher in EA (Lebiger-Vogel et al., 2022; Rickmeyer et al., 2017). Additional research indicates that temperamentally fearful children showed less of a stress response to fear-evoking film clips if their mothers were higher in EA (Gilissen et al., 2008). Finally, it seems important to make the point that when “EA” is being referred to as important, it is not always “sensitivity”. Given the resource investment in evaluations, it makes sense to use the entirety of the EA Scales/EA zones.

Daily activities of sleep and feeding/eating

Sleep

Bedtime maternal EA is associated with less infant distress and more sleep, coded through videosomonography, where the infant's sleep-wake cycles were objectively assessed (Philbrook and Teti, 2016a,b). More specifically, higher bedtime EA in combination with less close contact at bedtime was associated with better infant sleep, as if slight disengagement from interactions better prepares the infant for sleep. A noteworthy finding is that in an Israeli context, maternal sensitivity assessed during the day was unrelated to optimal sleep patterns (Scher, 2001), suggesting that the bedtime context is unique.

Feeding/eating (weight issues)

Atzaba-Poria et al., (2010) found both mother-child and father-child EA to be lower in a feeding disorders group. Gueron-Sela et al. (2011) reported that mothers who were more intrusive and less structuring were more likely to have infants with feeding disorders, potentially because of the worry about underweight. Further (Merras-Salmio et al., 2013), in mother-baby relationships, cow's milk allergy in the baby was associated with lower EA in the mother. The consistency in these findings might suggest that the parental side of EA affects the child's outcomes and yet, let us remember that the child's challenges with eating and the child's allergy may be impacting, even eliciting parenting behavior that is less than optimal. It takes two to tango! Additional research indicates that the parental side of EA (particularly higher sensitivity without optimal structuring) is related to picky eating in the child at 6 years of age (Steinsbekk et al., 2017), suggesting that being too empathic about the child's eating issues may potentially keep a child from trying new foods. Mother's intrusiveness during early childhood was associated with the child having higher body mass index at school age for both genders (i.e., overweight in the child) (Schlensog-Schuster et al., 2022). In contrast, maternal sensitivity during early childhood was predictive only of girls' body mass index. However, here, we are moving away from early childhood.

Prevention/interventions to alter EA

Numerous interventions have been tested to assess whether there are improvements in EA after the intervention. One such program is Parent Child Interaction Therapy, which appears to improve EA (e.g., Kohlhoff et al., 2021; Thomas and Zimmer-Gembeck, 2011). Video Feedback Intervention to Promote Positive Parenting (VIPP) was found to improve EA as well in adoptive children (Barone et al., 2018). Other interventions including Marte Meo (Høivik et al., 2018) and Circle of Security (Mothander et al., 2017) also have been found to improve EA. Not surprisingly, the Nurse Family Partnership (NFP) also has been found to alter EA (Robinson et al., 1997). PedaSens is a program out of Finland designed to alter the teacher's EA and has been successful as well (Harkoma et al., 2021). Less well known but highly significant clinical applications (tackling war trauma in the mother) also have improved EA (Pinchover and Shulman, 2019), primarily by focusing on playful ways of interacting. Many other programs are summarized in Biringen et al. (2014, 2022, 2023). A quick mention of the EA Briefing is in order here as it was designed specifically to alter EA, and has been tested in three studies so far, with children 0–3 years and their mothers (McConnell et al., 2020), with children 0–17 years and their mothers (Russell et al., 2025), as well as with children 0–17 years with children and their fathers (Lincoln et al., 2023). Results were positive. Few prevention/intervention studies that have been published yielded null findings, suggesting that EA is a malleable quality and resources can be and should be invested in helping parents become more emotionally available to their children.

Next steps

EA is one of the most investigated constructs and measurements in the field—it is a gold standard (Lotzin et al.,2015). As mentioned, as of today, there are close to 500 research publications, and, to our knowledge, more than any other observational tool, to date. Further, EA has been investigated in almost all continents, and in most sub-cultures in the US. Yet, while a popular tool across the globe, EA is not necessarily “popular” within each country. The main reason for this is that it is not quick and dirty and there is an associated time and cost investment involved. Many want to use self-report measures, yet there is only one published account of the link between the EA Self Report and observational EA (Vliegen et al., 2009). To make EA more popular within each country, including sustained use after publication and within practitioner communities, we are making concerted efforts to create EA hubs, including trainers in some countries.

Brief evidence-based prevention/intervention efforts that are available to all who may benefit

The EA field might benefit from investigating and utilizing prevention/intervention programs that are easier to implement and that can reach more people. The videoplayback programs are particularly useful to parents, and having more programs of this ilk be available to parents in community agencies should be a priority. Specifically, the EA Briefing is a single session of feedback and coaching, which means that it can be implemented in locations where mothers/children may pass through (e.g., domestic violence or homeless shelters) or added to traditional talk therapy sessions, even when the therapist is focused on treatment of the adult who also happens to be a parent. Marte Meo method developed by Aarts (2000) and empirically evaluated in terms of EA by Høivik et al. (2018) is also flexible and should be available to more parents. Similarly, Circle of Security has been incorporated (in brief forms) into centers focused on infant mental health, with now documented improvements in EA (e.g., Mothander et al., 2017).

The inclusion of fathers in research on high-risk environments and psychopathology

A great deal of work has been conducted on a mother's history of abuse or neglect and its sobering effects on not only mother's neural processing and neural activation (e.g., amygdala), but also mother-child EA (e.g., Olsavsky et al., 2019; Schneider et al., 2021). In addition, maternal exposure to the trauma of war also has an impact on mother-child interactions as measured by the EA (Van den Dries et al., 2012). Significant work on maternal mood disorders, including unipolar depression, bipolar depression, anxiety disorders, including with medication use and remitted disorders (e.g., Aran et al., 2021a,b) have been conducted. Bipolar disorders or bipolar features have also been explored (Trupe et al., 2018). A great deal of work on maternal substance use also has been conducted (Flykt et al., 2022; Porreca et al., 2018, 2025). Beginning work on maternal psychosis and EA is underway (Porreca et al., 2014).

Notably scant in this type of research on high-risk environments and psychopathology is fathers and their role in the family. Granted, basic research including fathers as participants is slower enrollment than research solely on mothers. Yet, basic science work on serious forms of psychological disturbance and prevention/intervention efforts to ameliorate emotional availability in the family system need also to include fathers, as the means to modify the father's style of interactions should not be just through the mother who has received the prevention/intervention efforts. Direct efforts that include mothers as well as fathers (or, at least other support people in the family) may create greater impact for infants and young children.

The connection of EA work with machine learning/artificial intelligence

AI can do amazing frequency counts of attachment behaviors (e.g., proximity seeking and proximity avoidance), varied facial expressions, and acoustic features that are suggestive of positivity and negativity. AI-powered facial recognition systems can use computer vision to analyze even micro-expressions that are missed by humans. AI can interpret gestures, such as hand movements and body posture (e.g., Salo et al., 2025; Stamate et al., 2023). It can also do voice and speech analysis, such as tone, pitch, and intonation to infer the emotional states of individuals. Further, physiological sensing such as heart rate, skin conductance, and body temperature can assist in detecting stress.

Can frequency of touch, acoustic features, and facial mapping software have a place in formulating summaries for a trained human observer, who then weighs in with the uniquely human “gut feeling” to arrive at an overall evaluation of EA? At this stage of AI, it seems prudent to say that the greatest barrier to AI entirely replacing human-centered observation is that the parent makes judgments about what to do (and how to connect) from prior experience with the baby or child, what works for this particular little person. AI is unlikely to be as human as a mother or father in knowing when they have connected, and the experienced human observer can more likely detect connection when it sees it. Similarly, while AI is likely to be very helpful to an interventionist in highlighting specific qualities that may be amendable to change, it is unlikely to be able to foster a connection with another human that helps to create the very mechanism for change – i.e., an empathic and trusting human relationship.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Maria Licata-Dandel, Charlotte Fresenius Hochschule, Germany

Reviewed by: Daniela Traficante, Catholic University of the Sacred Heart, Italy