In light of the above, the aim of the current study was to explore the effects of smartphone use and smartphone distractions on maternal attention and physiological function during two types of interactions with her infant: breastfeeding and face-to-face interaction. First, in line with previous studies, we hypothesized that the mother’s attentiveness, indexed by her gaze direction toward the infant, will be reduced during moments of smartphone use compared to smartphone unavailability. We expected this effect to be more pronounced during breastfeeding compared to face-to-face interactions–as breastfeeding is a setting in which mothers’ can more easily disengage attention to their infants. As this is the first study to assess changes in EDA and CO during breastfeeding and face-to-face interactions with regards to the smartphone, we aimed to explore if physiological function in these indices will be different during the smartphone use versus smartphone unavailability stages and in breastfeeding versus face-to-face interactions. Finally, in line with previous studies, we hypothesized that higher maternal attention towards the smartphone will be related to a blunted maternal physiological arousal (similar to an addiction like state), as well as to increased reported smartphone addiction.

Twenty mothers and their 3–6 months old infants (8 boys; 12 girls—65% were firstborns, 30% were born second, and 5% were born third) took part in the experiment. Criteria for study participation were mothers’ normal and corrected vision by contact lenses, and infant age was limited to three to six months. None of the mothers reported smoking cigarettes. Table 1 presents demographic information.

Missing Data. One mother’s data could not be collected for technical reasons, therfore, she was not included in the analyses. In three other cases, physiological data could not be fully collected during the study because the infant was crying/the mothers needed to discontinue breastfeeding, or due to electrode displacement. During breastfeeding, under the conditions of smartphone use/bag, physiological data were obtained in full from 19 participants. During the mute condition, physiological data were obtained in full from 16 mothers. During the face-to-face interactions, while the smartphone was on mute, data was collected fully from 18 participants. A total of 16 had full data for the entire study.

Before arriving at the lab, mothers completed several online questionnaires at home regarding their smartphone use, perceptions of their bond with the infant, ratings of their infant’s temperament, and demographic information. Upon arrival at the lab, mothers reported their baseline situational anxiety via a short questionnaire [65].

Participants were instructed to arrive at the lab session well hydrated and avoid caffeine/cigarettes two hours prior to the study. Mothers were physiologically monitored via a MindWare Mobile recording device (MindWare Technologies, Gahanna, OH) (See details on physiological collection and analysis below). Infants were not physiologically monitored during the study. After participants were connected to the physiological monitoring system, a five minutes baseline commenced, in which mothers were instructed to sit still and relax while their infant was under the care of a research assistant in a separate adjacent room. Following the baseline, mobile eye-tracking glasses were fitted to the mothers in order to measure their gaze patterns throughout all following experimental conditions. The entire lab visit was videotaped from three angles allowing full visibility of both interacting mother and infant. Video recordings, physiological recording, and SMI eye-tracking glasses were fully synchronized in time.

The experiment was comprised of three conditions in two phases. The phases were breastfeeding and face-to-face interaction, starting with breastfeeding. The conditions were: 1) Smartphone use 2) Smartphone sound on but unavailable to mothers 3) Smartphone on mute and unavailable to mothers. The experiment started with participants being asked to place their smartphone (sound on) on the table next to them and with the assistance of the experimenter, to wear the eye-tracking glasses. Next, participants were instructed to take their baby in their arms and engage in breastfeeding for five minutes. During this condition, five WhatsApp messages were sent to the participants’ smartphone by the experimenter (sitting in an adjacent control room). Some text messages with questions that mothers were required to answer (e.g., "what brought you to participate in our study?") and some messages with further instructions for the following conditions as well as a short situational anxiety questionnaire identical to the ones completed at the arrival to the lab ("well done! Now I would like you to answer the questionnaire attached in the following link"). In the next condition (breastfeeding while smartphone in the bag), participants were asked to place their phone in their bag, sound on, and were instructed not to respond to any received alerts (the experimenter entered the room before the measurement began and assisted if needed) while breastfeeding for another five minutes. In this condition, four messages were sent by the experimenter, and the mothers’ only heard the notifications. In the final condition of the breastfeeding phase (breastfeeding/smartphone mute), mothers were asked to mute their phone and to place it back in their bags for another five minutes. Participants then were given the time to end breastfeeding, as they saw fit. The following phase of the face-to-face interaction was conducted similarly to the previous one, with the three phone conditions–each lasting five minutes. Mothers were asked to have a face-to-face interaction with their infants as they normally do ("you can use your baby’s toy or the puppets in the room as well as one of the books"). The infants were placed in a baby swing across from the mothers, who were instructed to remain sitting during the measurement and to avoid standing or taking their baby out of the swing. WhatsApp messages were then sent to the participants’ smartphones in the same manner as in the breastfeeding phase. Before each condition ended, the participants were asked to complete the situational anxiety questionnaire that was sent again via a text link. At the end of the measurement of the conditions in which the phone was not available, research assistant entered the room and instructed them orally to answer the questionnaire in the link. After the final task, there was a final five-minute baseline recording, identical to the first. Finally, monitoring equipment was then removed. Participants were escorted to a control room where they were asked to watch specific video clips from the six different experiment condition*phases that were just recorded, and to answer some questions on their emotions when watching the clips–valence and arousal (the Affect Grid) [66] (Fig 1). Data from the Affect analysis is not reported here. Mothers and infants were thanked and compensated for their participation (received a coupon for a free coffee and an onesie for the baby).

We assessed physiological measures of the sympathetic branch of the autonomic nervous system (SNS) in mothers during breastfeeding and face-to-face interaction while also using their phone or getting notifications without being able to answer. We compared these conditions to a state in which the phone was muted and unavailable during interactions, so it could not overtly distract mothers. More specifically, we measured two types of physiological activity denoting SNS function: 1) Measures of electrodermal activity (EDA), assessed from the skin of the palm. 2) Measures of cardiological impedance, assessed from the heart (See details below on acquisition). We calculated statistics for two specific physiological measures of EDA: Tonic skin conductance level (SCL) and tonic period, and for one specific measure of cardiological impedance: Cardiac output (CO). Tonic SCL represents the tonic level of electrical conductivity of the skin calculated by the entire amount of SCL minus the phasic activity. Tonic period represents the amount of time spent in non-responsive states and is calculated by the total period of the waveform after removing all phasic responses (SCRs). Cardiac output indicates how much blood is being pumped out of the heart at any given time in liters per minute and is calculated as Heart Rate (HR) times Stroke Volume (SV).

Maternal attention to the infant was operationalized via the assessment of mothers’ gaze patterns with an eye-tracking mobile device (Senso Motoric Instruments; Teltow, Germany, SMI, https://www.smivision.com)—a non-invasive and robust system designed to be used as a common pair of glasses (weight 75g). The SMI head-mounted system includes two small cameras on the rim of the glasses capturing the eye movements of the wearer, and a front view camera that captures the participant’s line of sight. The range of eye-tracking is 80° horizontal, 60° vertical with a binocular 60Hz temporal resolution and up to 0.1° spatial resolution. This is combined with a recording from a 24-Hz front view camera with a field of view: 60° horizontal and 46° vertical. Before the acquisition, for each participant, a three-point online calibration was conducted, followed by a check of the calibration’s accuracy. The calibration procedure included three dots on the wall at a distance of approximately 1 meter from the participant. Mothers with corrected vision (glasses or contact lenses) were instructed to use contact lenses in order for SMI glasses to be used.

The Smartphone Addiction Scale (SAS) was used to measure daily-life smartphone use (disturbance, positive anticipation, withdrawal, cyberspace-oriented relationship, overuse, and tolerance) [48]. The following surveys were collected but not included in the current report: The Infant Behavior Questionnaire–Revised Very Short Form (IBQ-R) measures infant’ temperament [71]; The Mobile Attachment Scale (MAS) [39]; The Maternal-to-Infant Bonding Scale (MIBS) [72]; The short form of the State-Trait Anxiety Inventory [65]; The Affect Grid [66].

In order to assess if experimental phases and conditions, as well as maternal gaze direction (AOI: towards her infant or her phone), were related to the amount of maternal fixation on each AOI, we conducted a two-factor ANOVA with repeated measures (with Bonferroni post hoc correction for multiple comparisons). The analysis revealed a significant interaction effect between experimental phases (breastfeeding versus face-to-face) and AOIs in predicting the amount of the normalized dwell time (ms/coverage) mothers’ directed to each AOI [F(1,17) = 80.42, p < .001, partial eta² = .83]. As can be seen in Fig 2, the normalized dwell time towards the smartphone was highest during breastfeeding (M = 3.41e+6, SD = 1.27e+6) and lowest during face-to-face interactions (M = 681,587, SD = 210,648). An opposite pattern was observed for the normalized dwell towards the infant–it was highest during face-to-face interactions (M = 3.19e+6, SD = 1.86e+6) and lowest during breastfeeding (M = 939,495, SD = 518,675) (Fig 2). No significant main effects were revealed for either the experimental phase [F(1,17) = .60, p = .45, partial eta² = .03] or for the AOIs [F(1,17) = .01, p < .95, partial eta² = .00].

In order to assess if experimental conditions and phases were related to maternal physiological activity, we performed three repeated measures ANOVAs, in which CO, tonic SCL, and tonic period were the dependent variables. Independent variables were the experimental phases (breastfeeding and face-to-face interaction) and conditions (smartphone use, smartphone in the bag sound on, smartphone in the bag on mute). Results of the first ANOVA with CO as the dependent variable revealed no interaction effect between the experimental phases and conditions [F(2,26) = .23, p = .80, partial eta² = .018]. A significant effect was found for the experimental phases [F(1,13) = 5.01, p = .04, partial eta² = .28], such that CO was higher during breastfeeding (M = 26.60, SD = 2.10) compared to face-to-face interactions (M = 23.00, SD = 2.10). No effect was reveled for the experimental conditions [F(2,26) = 0.19, p = .83, partial eta² = .01] (see Fig 3A).

We used Pearson’s correlations to assess associations between the reported level of smartphone addiction and physiological measures. During breastfeeding: We found that tonic SCL during smartphone use and tonic SCL when the smartphone was on mute were negatively related to addiction scores (r = -.56, p = .01; r = -.58, p = .02 respectively). While in face-to-face interactions: We found that tonic SCL when the smartphone was in the bag and tonic SCL when the smartphone was on mute were negatively related to addiction scores (r = -.55, p = .01; r = -.50, p = .04 respectively). Additionally, during face-to-face interactions tonic period when the smartphone was in the bag was positively related to smartphone addiction scores (r = .50, p = .03). CO was not significantly related to addiction scores in any of the study’s phases or conditions (See all correlations in S1–S3 Tables). Overall the results indicate that in many of the study’s phases and conditions, higher EDA was related to lower levels of mothers’ addiction to their smartphones.

In order to descriptively assess if there was a potential link between mothers’ shifting their gaze and attending to their smartphone and physiological arousal we looked at mothers’ skin conductance responses while getting notifications and having to read and respond to them. In Fig 4 we show two such examples from different mothers who participated in the study. As can be seen, at least descriptively, incoming messages (indicated by red vertical lines in the figures) were sometimes followed by a distinctive skin conductance response. These immediate responses may demonstrate that a shift in attention towards the smartphone was accompanied by a physiological arousal response (Fig 4).

The limitations of the current study are mainly the small sample size and its exploratory nature, which entail that further studies corroborate and expand these initial results. Moreover, in the current study, we did not assess immediate physiological responses to received notifications, and hence, future studies are required to explore breastfeeding mothers’ immediate physiological responses when smartphone distractions occur. Further studies are also needed to examine if our results are unique to the breastfeeding context or may appear in other feeding situations in non-breastfeeding mothers. Another way in which our research design could be strengthened is to include a baseline condition in which the mother is not interacting with her infant but instructed to be involved with her smartphone “as she usually does”. This condition may reveal baseline differences in maternal proclivity towards checking her cell phone and could further develop the findings of the current study. An exploratory descriptive analysis of several mothers’ EDA time-series’ indicated there may be “incoming message”- specific physiological responses in EDA that accompany the shifts in attention from infant to smartphone. As we did not have a-priori hypotheses regarding these effects, we suggest they be further validated in future research. Finally, these results should be expanded in longitudinal research designs to examine the developmental influences of smartphone use on infants and the mother-infant dyad.

Notwithstanding our project’s strength is in its multimodal and interactive nature: A mental, behavioral, and physiological examination of mothers during real-life situations that involve interactions with their infant and smartphone use. To our knowledge, this is the first presentation of data and findings that characterize behavioral and physiological markers of a preference to the smartphone in mothers’ that is context specific—breastfeeding versus face-to-face interactions. Interestingly, some physiological responses were similar to those reported in the literature on substance addiction. Our results should be further explored to fully understand the outcomes of smartphone use on characteristics of the mother, the infant and their bond in this digital era.