Depressed Mothers' Touching on Their Mfants' Behavior Were Investigated During the Still-Fece Situation

Depressed Motliers' Touching Increases Infants' Positive Affect and Attention in Still-Face Interactions

Martha Pelaez-Nogueras Florida International University Tiffany M. Field University of Miami

Ziarat Hossain Fort Lewis College Jeffrey Plekens James Madison University

PELAEZ-NOGUERAS, MAHTHA. FIELD, TIFFANY M., HOSSAIN, ZL^RAT, and PICKENS, JEFFREY De- pressed Mothers' Touching Increases Infants' Positive Affect and Attention in Still-Face Interac- tions. CHILD DEVELOP^^i:NT, 1996, 67, 1T80-1792. The effects of depressed mothers' touching on their mfants' behavior were investigated during the still-fece situation. 48 depressed and nondepressed mothers and their 3-month-old infants were randomly assigned to control and experimental conditions. 4 successive 90-sec periods were implemented: (A) normal play, (B) still-face-no-touch, (C) still-face-with-touch, and (A) normal play. Depressed and nondepressed motliers were instructed and shown how to provide touch for their infants during the still-face- with-touch period. Different affective and attentive responses ofthe infants of depressed versus tie infants of nondepressed mothers were observed. Infants of depressed mothers showed more positive affect (sniiles and vocalizations) and gazed more at their mothers' hands during the still-face-with-touch period than the infants of Bondepiessed mothers, who grimaced, cried, and gazed away from their mothers' faces more often. The results suggest that by pro\'iding touch stimulation for their infants, the depressed mothers can increase infant positive affect and attention and, in this way, compensate for negative effects often resulting from their t>'pical lack of affectivity (Hat facial and vocal expressions) during interactions

Early interaction disturbances place in- nal depression on early infant interactions fants of depressed mothers at risk for later and development and have identified the affective and socioemotional disorders behavior patterns of depressed mothers as (Field, 1992; Gaensbauer, Harmon, Cytryn, unresponsive, insensitive, ineffective, non- & McKnew, 1984; Zahn-Waxler, Cummings, contingent, emotionally flat, negative, disen- McKnew, & Radke-Yarrow, 1984). Having a gaged, intrusive, avoidant of confrontation, depressed mother increases by three times a and generally less competent and unin- child's risk of developing the abnormalities volved with their infants (e.g., Campbell, characteristic of depressed mothers (Weiss- Cohn, & Meyers, 1995; Cohn, Matias, Tron- man et al., 1984). Numerous studies have ick, Connell, & Lyons-Ruth, 1986; Cohn & documented the negative impact of mater- Tronick, 1983; Field, 1984, 1986; Lyons-

The authors wish to thank Julie Malphurs, Jeanette Gonzalez, Claudia Larram, and Angie Gonzalez, for aiding with data collection and Roberto Pelaez for help with data analyses. The research reported in this article was supported by National Institute of Mental Health Research Scientist Award MH00331 and National Institute of Mental Health Basic Research Grant MH46586. Portions of these data were presented m March 1993 at the meeting ofthe Societ>' for Research in Ghild Development New Orleans, Louisiana. Gorrespondence and reprint requests should be addressed to Dr. Martha Pelaez-Xogueras, Department of Educational Psychology and Special Education, Florida International Universit\', Miami, Fl 33199. [ChiUDevelopment, 1996,67,1780-1792. © 1996 by the Society for Research in Child Development, Inc. All rights reserved 0009-3920/96/6704-0031$0100]' Pelaez-Nogueras et al, 1781 Ruth, Zoll, Connell, & Grunebaum, 1986; Researchers have prospectively studied Pelaez-Nogueras, Field, Cigales, Gonzalez, infants and toddlers of depressed mothers to & Clasky, 1994). analyze the processes and mechanisms whereby depression may affect infant be- Infants of depressed mothers, in turn, havior. Diverse mechanisms have been hy- appear to develop a depressed mood style as pothesized to produce the negative out- early as 3 months. The "depressed" infants comes observed in infants and children of typically exhibit less attentiveness, fewer depressed mothers (e.g., Reardslee, Bemp- smiles, more fussiness, more gazing away, orad, Keller, & Herman, 1983; Cummings & and lower activity' levels when interacting Cicchetti, 1990; Hammen, 1992). However, with their depressed mothers than infants of elucidation of the mechanisms and pro- nondepressed mothers (Cohn, Campbell, cesses involved in the transmission of soci- Matias, & Hopkins, 1990; Gelfand & Teti, oemotional behavioral problems from de- 1990; Goodman, 1992). Moreover, maternal pressed mothers to their infants is still a depression has been significantly associated major challenge for developmental research- with attachment insecurity among infants ers. This is because early development of and preschoolers (Teti, Gelfand, Messinger, infant depression may result from the inter- & Isabella, 1995). Infants of depressed moth- action of multiple influences, including bio- ers, however, do not necessarily generalize logical factors and psychosocial factors. Sev- their "depressed mood" to other adults. eral models of early development of WTien the infants of depressed mothers in- depression have been proposed, including teracted with their nondepressed nursery mutual regulation (Tronick & Gianino, teachers, the infants' behavior rescovered, 1986), multivariate cumulati\'e risk (Field, and their activity levels and positive affect 1992), and temperament and genetic predis- rates were higher than when interacting position (\\Tiiffen & Gottlib, 1989). These with their depressed mothers (Pelaez- models have focused on the effects of multi- Nogueras et al., 1994). ple factors that include prenatal influences In general, depressed mothers and their and postnatal experiences. An infant show- infants appear to share their behavior states, ing a depressed-mood pattern could be spending more time in negative attentive/ (a) biologically predisposed to depression affective behavior states than pondepressed due to prenatal exposure to the depressed mQther-infant dyads (Field, Healy, Gold- mother's physiological imbalance and hor- stein, & Guthertz, 1990). Different profiles monal status or due to a potentially congeni- of behavipr have been identified, including tal predisposition or (fc) environmentally af- disengaged mothers (withdrawn and pas- fected due to continuing maternal depressed sive) and intrusive mothers (e.g., angry facial behavior patterns, like unresponsiveness expressions and intrusive poking of the in- and flat affect Thus, multiple factors seem fant) (Field et aL, 1990; Malphurs, Raag, to be affecting both mothers' depression and Field, Pickens, & Pelaez-Nogueras, 1996). infants' behavioral patterns denoting "de- But despite the variability observed in the pression," and there are multiple interac- mothers' interaction stj-Ies, the infants of tions between these underlying affective, both disengaged and intrusive mothers are perceptual, physiological, and biochemical usually uniformly distressed. Also, whether processes. experiencing postpartum or chronic depression (Campbell et al., 1995), the common The objective of the present study was findinia in the literature is that the depressed to determine whether depressed mothers mothers' negative mood states and lack of can improve their infants' attentive and af- affective responses negatively affect the fective responses by providing touch stimu- child's behavior. In this way, the infants of lation during still face interactions. Touch, depressed mothers begin to show growth as a source of stimulation, has received little and developmental delays at 1 year if their attention in the mother-infant interaction lit- mothers remain depressed over the first erature. The studies reported below suggest year. Normally, the developmental delays that tactile stimulation is a significant con- are manifested by inferior performance on tributor to infant growth and social devel- Bayley Mental and Motor Scales at 1 year of opment. age, but other behavioral deficits have also Touch Stimulation been noted, including heightened emotion- in High Risk Populations ality and a lower level of s>Tmbolic play Studies involving touch of premature in- (Field, 1984; Gaensbauer etal., 1984; Samer'- fants and neonates have reported improve- off & Seifer, 1983; Whiffen & Gottlib, 1989). ments in physiological growth, motor/refiex. 1782 Child Development

For example, an Inter^Sn to study mother-mLl

i?Si 5iSnsi ^^g with negative affect and other (Scafidietal.,19l)) pp behaviors. Stock and Muir (19901 p during Face-to-Face Interactiom "** '^**'" mothers were asked to be' Face-tp-face interactions are a primary ^responsive, sifent, and not to way^behavior disorders seem to be transmit-- *f»"*^ during the stfll-face epi- ted from motlier to infant (Cohn et al. 1986- displayed more grimacing and fq«m' 'T^*'^^"^*?'"' ^'''*^*' ^ Goldstein: ^^^^^ ^ perms o£ normal 1986). The qualitj- of infant behavior has , when touch was intro- been,related to the wiresponsiveness and ^"^ still-face period, infants" emotional unavailability of their mothers and attention was higher. It ?^""S*'^^ein1:emctions(Samerbff"S^*'^^ein1:emctions(Samerbff& SeifeSifr^ be deteimined, however, 13; Tronicfc & Gianino, 1986). Only a few *« of depressted mothers are studies have investigated the, effects of ma- sensitive to maternal touch than in- ternal toucji during faee-to-face mteractions o^'Nondepressed mothers. As yet, no witli th.eir infants or ike use of touch inter- investigated the effects of touch ventions to faGilif^te better interactions be- d mothers using the still-face tween mothers and tlieir infknts (e.g. Field

^™»«^ the negative effects elicited by their for example, Pelaez-Nogueras et al (in t f ^ providing additional touch for ps^) found that touch can reinforce and We thought it important t» ex- maintain high rates of infant eye contact re i^'°'H^^^~ depressed moth- sponses, vocalizations, and smiles during ^^'^ '^^^ts, to regulate affect face-to-face inteTactions. In that study usine increase their attention. The a ^tjiKhronous-reinforcementopem7tn^ce- ht ifants of depressed moth- distressed as the infants dure, touch stirmilition (gentle rubbtogof f the infant's arms, legs, and feetfwasnro mothers during the stilJ- vided^ by a care^ver *hile the nfent w^s" ,îta^tio« because they were mîng eye conSet with hL Afêr stS &eir mothers with flatWt ponditioning sessions, the infants showed / ^" ^^^^^ intote, touch preferences for the reinforcing stimukHon (or compensate for) the lack th^t included touch, as showf by^h^S >^^ the other sources (i.e.. that they smiled and vocalised US and ^ oth^r hand, the u. made more eye contact with the caregiver mothers were ex- Thos^ findingssugges t that infkxts'atS^ and positive affect can be leinforce^d and ced in a stjU-face- maiatained by an adult providing ? period, because dieir mothers' tactile stimulLion during feS £ *''™' ^« ««expeeted and actions m experience.

by he, adopt a .tation^ ™od,er Pelaez-Nogueras et al. 1783 positively than infants of nondepressed instruments in research on nonclinically de- mothers when optimal touch (mild strokes/ pressed samples. This self-report scale was movements) was introduced in the still-face used rather than a diagnostic interview be- situation. Thus, maternal behavior was ma- cause Cohn and Campbell (1992) have re- nipulated, touch was standardized, and the ported that depressed mothers' interaction main grouping (independent) variable was behaviors are more highly correlated with maternal depression score. By standardizing self-report depression scores than they are touch we minimized differences in de- with diagnostic interview measures. Moth- pressed and nondepressed mothers' kinds of ers with BDI scores of 13 or greater (cutpoint touch and were able to assess whejther ma- of depression in most research protocols) ternal depression could account for the dif- were assigned to the depressed group and ferences in infant behavior. mothers with scores of 9 or less were assigned to the nondepressed group. We ad- Method ministered the BDI to 61 mothers to yield our sample of 24 depressed mothers. In pre- Subjects vious studies withi this population, approxi- Fort>'-eight 3-month-old infants (mean mately 30% ofthe mothers sampled received age = 13.5 weeks, SD = 1.2) and their moth- scores greater than 16 on the BDI (e.g., Eield ers (mean age = 19.1, SD = 2.7) partici- et al., 1990). The mean BDI score for all de- pated in this study. All infants were healthy, pressed mothers in our sample was 21 (SD born at gestational age (M = 38 weeks), = 9.1), ranged fram 13 to 52, and for the were of normal birthweight, and had no his- nondepressed moAers was 4.1 (SD = 2.7), tory of medical complications. Subjects were ranged from 1 to 9, Mothers with BDI scores recruited from a longitudinal study sample of zero, 10, 11, and 12 did not participate in of low socioeconomic status based on the this study. The BDIs were administered 15 two-factor Hollingshead Index. Mothers min before the interaction in a waiting room were primiparous black (33%), Hispanic next to the laborator>' by a research assistant. (40%), or Caucasian (7%), and were single (86%) adolescents, and their infants were normal full term infants. Three mother- Apparatus and setting.—Infants were infant dyads needed to be rescheduled be- seated, in an infant seat facing their mothers cause the babies were fussy and sleepy. at a distance of approximately 15 inches. Mothers were seated directly facing their in- Mother-infant dyads were assigned to fants at eye level. Two cameras, located on one of four groups: Depressed Mothers- either side of the mother-infant dyad, were Experimental (N = 16), Nondepressed connected to a video recorder and a special Mothers-Experimental {N = 16), and De- effects generator to yield a split-screen im- pressed Control and Nondepressed Control age. One camera recorded the frontal view (A^ = 16). The Beck Depression Inventorv- of the infant, and the second camera re- (BDI) scores defined the depressed and the corded the mother's face and hands. A time- nondepressed groups (depression classifica- date generator connected to the monitor was tion is described in detail in the fiDllowing used to time the duration (in minutes, sec- section). The groups did not differ on demo- onds, and milliseconds) of each period for graphic variables, including age, ethnicit\', subsequent coding. maritaf status, and SES, resulting in a homogeneous sample. To ensure group equiva- Design.—A repeated-measures be- lence, infants and their mothers were as- tween-groups design was implemented: two signed to control or experimental groups groups (depressed vs. nondepressed) x two through a random stratification procedure, conditions (control vs. experimental) x four stratifying in accordance to maternal depres- successive periods: (A) 90-sec normal inter- sion score. action, followed by (B) 90-sec still-face-no- touch, (C) 90-sec still-face-with-touch, and Procedure finally, (A) 90-sec normal interaction. Six- Maternal depression assessment.— teen additional mother-infant dyads were Assignment to the depressed groups was used as a no-still-face control group. Mothers based on cutoff scores on the Beck Depres- in the control condition only received the sion Inventory (BDI; Beck, Ward, Mendel- normal interaction instructions across the son, Mach, & Erbaugh, 1961). The 21 BDI four consecutive periods of the study. The it^ms are scored on a fbur-point scale indi- design compares the 16 controls (half de- cating absence/presence and severily of depressed and half nondepressed) to tlie 32 ex- pressed feelings, behaviors, and symptoms. perimental (half depressed and half nonde- The scale is among the commonly employed pressed). The order of the periods was not 1784 Child Development Aan 15 sec, Ae session was interrupted and counterbalanced because the purpose was to rescheduled. A total of five mother-mfant dy- have a still-face-no-touch period preceding ads needed to be retrained and rescheduled a still-face-wiA-touch period specifically to for a second visit. induce distress in order to increase the chances of getting an effect and thus to com- Touch procedure.—]ust before the still- pare Ae depressed and nondepressed dy- t-ace-wiA-touch period, ali moAers m Ae ads' performances. experimental condition received a ôriet demonstration of optimal touch The opti- lnstmctions.~-The total procedure re- mal" touch procedure involved a mother quired approximately 8 mm. During Ae four stroking and rubbing rhythmically the m- 30-sec intervals between Ae four period, all fants' arms, legs, and feet using the five fin- mothers were given instructions. Iristruc- gers of boA hands for Ae duration of Ae tions were standard for all mothers. To ad- still-face period (90 sec). The experimenter dress Ae question of wheAer infants oi de- modeled "gentle pressure in slow circaiar pressed and nondepressed mothers respond ir'otions at a rate of approximately one circu- diffferentially to touch when their motliers lar rub per sec. Negative touch was avoided. pose a still face, it was important to reduce Negative touch involves rough tickling, pok- variability in moAers' behavior during Ae inA^d tugging while interacting wiA Ae still-face 'situation. For Ais reason, we irp- infant, including poking Ae baby s face, posed still-face instructions to all mothers in ar'ms, or stomach, or pinching or sqiieemng Ae experimental condition. Ae infant, or pulling or shaking Ae infant. MoAers were instrticted not to tickle or Before Ae first normal play periodôf in- poke, their infants during Ais procedure, nor teraction^ mothers in the efPenmental con- to pull intensively their infants' legs or an^s. dition were instmcted to play vviA then m- The motliers' touch was checked routinely fan*s as Aey would normally do at home. darin« the interactioBS to make sure Aey For the second period (stiU-face-no-touch), were providing toiieh as instnieted. Aese mothers were instructed to look/gaze at Aeir' infants wiA a neutral expression Behavior coding,—The onset and offset and to refrain frorn speaking, smilmg, and of the videotaped behavior were registered toucliing the infant during this period tor bv pressing numeric codes on a laptop com- Ae third period of interaction (still-iace- DUter All behavior modalities were coded with-touch) instructions were given to looK/ separately. The behaviors were coded con- gaze at Ae iniknt wiA a neutral expression, tiliouslv and featured a second-oy-second to refi-ain from speaking aad smAng, but to listing of behaviors and a matrix «f percent- touch Ae in&fltas modeled. In Ae last nor- age ttme Ae behaviors occurred iGuAeite mal period, mothers received the same m- fe'Reia. 1989). One view of &e videorecord structiions as in the first normal period. wa. used per each modality: (1) mfantfcml To ensure Aat moAers maintained a eKpressions (three codes: smile, «eutral gr^ still face Aroixghout Ae still-face pemds mace) (2> infant vocal expressions (tlijree continuous monitoring was conducted by Sries^ positive voeafertions, BO vocahâ- Ae second research assistant observing the tiom, a4 p«>tBst/cryir«), (3) iniaBt gaze be- interaction from Ae observation room, ihe havior (used three codes: gaze at moAers obser^r constantly checked that moAers fece, SBze away fiom moAer's fece, gaze at were complying wlA instructions and were SotheVs Wnds). Thus, coding of these mea- not making anj-change in facial expressions, sures required three separate viewu^s of each record. In this way, coding o. the m- SS'tehavior. i-ciudBd three posrtiva be. smiling, (2) vocalizing.. (3) gazmg exs' ha3»ds, and Areejaegative at £c^ behaviors: (4) crying, (5) grimacmg, as instjucted. All mothers in both groups and (6) gazing away from mom. coniplied with instructions (> 90^°^^^*^ time) In those cases where mothers were ForJnfant smiling to be coded^the infmt EoUowing instructions and smiled, _wcal- niouA had to be "i^Ptumed wheAer Ae I or touched Aeir iniants incorreetiy the moutti was open or closed. For lijtant gri ' was interrupted and Bostponed tor a S°Ss'the iSanfs mouA had to be toned ts or curled or Ae Infant had to be cry- For gazing away from Ae moAer, Ae i h J to fe looking at any oAer place Pelaez-Nogueras et al. 1785 but the mother's face, hands, or body. Posi- Results tive vocalizations were discrete sounds like The first analyses were a 2 (group: de- those involved in cooing and babbling (but pressed vs. nondepressed) X 2 (condition: the infant could not be fussing or protesting). experimental vs. control) x 4 (periods of in- For crying, the infant had to be grimacing teraction) MAKOVAs on infants' positive be- and emitting nondiscrete/loud sounds. haviors (smiling, vocalizations, and gazing at mothers' hands), and on infant's negative be- Given the highly standardized proce- haviors (grimacing, cr>'ing, and gaze away dure of this study, for control purposes the from mothers). For the first MAKOVA on in- mother's behaviors were also coded, that is, fant positive behaviors, the analyses yielded mother's touch, facial expressions (smiles, a significant three-way interaction effect of negative/angry, neutral), and mother's vocal group X condition x periods, F(9, 36) = sounds were coded. This allowed ns to en- 2..56, p < .05. Then, significant main effects sure that mothers were following the still- were also observed for group, F(3, 42) = face, no voice, and no-touch and touch in- 4.28, p < .01, and condition, F(3, 42) = 2.7, structions. For touch behavior, five p = .05. For the MANOVA on negative be- (numeric) codes on the laptop computer haviors, the analyses yielded a significant were used to code touch behavior: (1) three-way interaction effect of group x con- mother's hand resting on baby, (2) mild dition X periods, F(9, 36) = 2.56, p < .05. touching (stroking, caressing, rubbing), (3) For the negative infant behaviors significant intense touching (tickling, poking), (4) mild main effects were also observed for group, movement (lifting baby's feet or arms in F(3, 42) = 3.90, p < .05, and for condition, slow, rhythmic cycling), and (5) intense F(3. 42) = 2.68, p = .05. movement (cinick intense movements of arms and legs or pulling arms or legs) (Stack Separate analyses for the control and the & Muir, 1990). The purpose to measure experimental conditions revealed: (1) no touch was to ensure that mothers were pro- changes in the control condition on any be- viding mild touch and movements (2 and 4) havior were obse]"\'ed over time, across the for at least 75% of the time during the still- four periods; (2) no significant differences in face touch period and were not making in- the behavior of the infants of depressed and tense movements or pulling the infant's legs nondepressed mothers were observed in the or arms. control condition across the four periods; (3) no differences were obserx'ed between Observer Reliability the control and experimental mother-infant Observers were unaware of the hypoth- dyads in the Hrst normal period. These anal- eses and of the mothers' depression status. yses suggested that tlie control and experi- The two independent raters were trained to mental conditions were similar at the begin- 90% reliability on each response category ning (first normal period) of the study and with an experienced rater. Reliabilit\' of the that the infants were Bot fatigued over time. behavior measures was determined on one- third of the sample. Product-moment corre- A significant main effect of group (de- lation coefficients were obtained on the per- pressed vs. nondepressed), F(3, 28) = 7.42, centage scores of primary and secondary' p <.OO1; a significant group x periods inter- obser/ers on all response measures of infant action, F(9, 22) = 4.77, p < .001, across pe- and mother beha'viors. Observer reliability, riods for positive infant behaviors; a main calculated separately for each response mea- effect of group, F(3, 28) = 2.72, p =.06; and sure, was at p < .001 for each measure. The a group X periods interaction effect, F(9, 22) reliability coefficients obtained for infants' = 3.02, p < .01, for negative infant behav- behaviors were as follows: infant smile, r = iors, were each obtained in the experimental .96; infant vocalization, r = .92; infant gaze condition. at hands, r = .90; infant grimacing, r = .97; The repeated-measure ANOVAs for in- infant crying, r = .94; infant gaze away, dividual dependent measures in the experi- r = .98. For mothers' behaviors the reliabil- mental condition were conducted to assess ity coefficients were: (1) mother's hand rest- for main and interaction effects associated ing ort baby, r = .99; (2) mild touching, r = with the primary grouping variable (de- .95; (3) intense touching, r = .92; (4) mild pressed vs. nondepressed) across the four movement, r = .96; and (5) intense move- periods (normal, still-face-with-touch, still- ment, r = .93; vocal sounds, r = .98; smiles, face-no-touch, normal). Simple effects analy- r = ,92; negative/angry face, r = 88; and ses between depressed and nondepressed neutral face, r = .96. scores were conducted only when the re- 1786 Child De^êlopment peated measures showed significant interac- often than infants of nondepressed mothers tion effects (Winer, 1971). Paired f tests were during the first nornial episode, F(l, 30) = also conducted Within subjects to compare 5.10, p < .05, still-face-with-touch period, the between still-face-with-tpuch and still- F(l, 30) = 4.97, p < .05, and dming the last face-no-touch periods (shown by subscripts normal period, F(l, 30) = 1.3.30, p < .001. in Table 1). The results on each variable follow. Infant Negative Behaviors Grimacing.—An ANOVA on grimacing Infant Positive Beliaviors yielded a group effect (depressed vs. nonde- Smiling.—ANOVA results for smiling pressed), F(l, 30) = 5,50, p < .05, and a yielded a group trend (dejpressed vs. nonde- group X period interaction effect, F(3, 90) pressed), F(l, 30) = 3.68, p < ;06, and a sig- = 7:01, p < .001. For the depressed group, nificant group X period interaction effect, infant grimacing decreased from the still- F(3, 90) = 2.94, p < .05. The propoition of face-no~touch period tq the stiil-face-with- smiling decreased from the normal period of touch period in the depressed group only, interaction to the still-face-no-touch period i(15) = 2.58, p < .03. simple main effects in both groups (Table 1). However, only the analysis revealed that infants in the de- depressed group showed a significant in- pressed group grimaced less often than the crease in smiling from the still-face-no-touch infants of nqsndepressed mothers during the to the inimediately following still-face-wilh- still-faee-with-toiich period, F(l, 30) = touch period, til5) = -2.33, p < .05. Simple 11.15, p < .005. Grimacing was also less fre- main effect tests performed on infant smiling quent, F(l, 30) = 7.62, p < .01, in the de- revealed that the depressed and nonde- presseti group compared to the nondepressed groups differed in the still-face- pressed group during the last normal period. with-touch period, F(l, 30) = 11.15, p < ,005, and in the last normal period, F(l, .30) Crying.—For crying, only a group x pe- = 16M0, p < .005, with the depressed giroup riod j'nteraction effect was obtained, F(3, 90) smiling more. = 2\B2, p < .05. For the depressed group only, infant cmng decreased from the sMll- face-iiO"toueh period to the still-face-with- Voeâlizations.—An ANOVA yielded a touch period, t(15) = 3.43, p < ,005. Crying significant group effect (depressed x"^. non- was lower, F(l, 30) = 4.98, p < ,05, in the depressed), F(l, 30) = 4.79, p < .05, and a depressed group cOmpaired to the nonde- significant group x period interaction effect, pressed group during the still-face-with- F(3, 90) = 8.74, p < .001. The proportion of touch period. Crying continued to be lower, time spent vocalizing decreased from the F(l, 30) = 4.39, p < .05, fbr the depressed first normal period of interaction to the still- group compared to the nondepressed group face-without7touch period for the nonde- duriflg" the last norrnal period. pressed group only (Table 1). Post hoc simple main effects revealed group differences Gazing away from mother.—An in vocalizing during the still-face-with-touch ANOVA yielded a group effect (depressed period, F(l, 30) = 5.52, p < .0.5, and during vs. nondepressed), F(l, .30) = 5,79, p < ,05, the last normal period, F(l, 30) = 12.50, and a group x period interaction effect, F(l, p < .001, witii infants of depressed mothers 30) = 7.55, p < .001. The proportion of time vocalizing m:ore than infants of nonde- the Inlaats gazed away from their mothers pressed mathers;. The differences noted in significantly decreased, tfl5) = 5.65, p < infants' vocalizations in the normal periods .001, ffepi the still-faee-no-fouch to the im- werd not signiftcant. mediatipiy following stili-face-with-touch period ip the depressed groiip but not in the teasing at mother's hands.—Results re- nondepressed group. Simple main effects vealed a main effect for group (depressed vs. anai|fsi| gieifbrmed oii giaing away from the nondepressed), F(l, 30) = 10.90, p < .005, motfifit revealed that &e nondepressed and a group x period interaction effect, F(3, group gaaed away more than the depressed 90) = 3.72, p < .01, in gaze at mouther's groul diiring the stiU-fece-with-touch pe- hands. As expected, the proportion of time riad,:-'Fp-, '30> =18,00, p < .001, and also infapts gazed at mother's hands increased dujriiiig the last normal period, F(l, 30) = sigiiificantly from the stili-face-no-touch pe- 19,l|,-p < ,005, The difference obsei-ved in riod to the still-faee-with-touch period for the f rst normal period was not significant. both the depressed group, t{15) = —4.88, p < .001, and for the nondepressed group, Mothers' Behavior t{lS) = 5.53, p < .001. Simple main effects Repeated-measures MANOVA on ma- analysis revealed that infants of depressed ternal tottch revealed BO significant main ef- motiiers gazed at their mothers' hands more fects of" groups (depressed vs. nondepressed) 3 1788 Child Development or group X period interaction effects (p > however, the effects were more soothing for .10). Touch was provided almost continu- the infants of depressed mothers and gaze ously by all mothers (97.7% of the time by aversion significantly decreased, but only depressed mothers and 93.7% ofthe time by decreased for the infants of depressed moth- nondepressed mothers) during the 90-sec ers. In this way, the distress caused by ma- still-face-with-touch period (Table 2). The ternal lack of facial expressions and voice instructions provided to both groups for the was reduced by instructing mothers to ac- still-face-no-touch and for the still-face- tively touch their infants. These findings with-touch periods minimized siny potential suggest that the effects caused by the still difference in maternal behavior. Also, touch face (lack of emotional expressions) can be instructions seemed to produce a carryover partially eliminated (or reduced) by mothers effect of the mothers' touch behavior from actively touching their infants while still fa- the still-face-with-touch period to the last cially and verbally unresponsive. nornial period of interaction, during which In general, both groups of infants botla depressed and nondepressed mothers seemed to like touch, and they showed it by touched their infants more than during tlie smiling and vocalizing more when they first normal play period. The higher amount were touched. This study extends the previ- of touch during the last normal period com- ous findings (Pelaez-Nogueras et al,, 1996; pared to the first normal period could have Stacks Muir, 1990,1992) by examining dif- accounted for the significant differences ob- ferences between depressed and nonde- served in infant behawors between these pressed groups. In addition to finding that two periods. infants of depressed mothers smiled and vo- Overall MANOVA for mothers' vocal calized more, oriented more to their de- sounds, smiles, and stillface revealed no sig- pressed mothers, and cried and grimaced nificant main effects of groups (depressed vs. less than infants of nondepressed mothers nondepressed) or group X period interac- during the still-face-with-touch period, we tion effects (p > .10), This result was also found that during tlie final return to normal expected given that the maternal behavior play period infents of nondepressed mothers in both groups was highly standardized and did not appear to "recover" from the dis- urlder experimental control during the still- tressing still-faee periods, and they began to face-no-touch and still-face-with-touch pe- cry, grimace more, gaze away more, and to riods. smile and vocalize less compared to infants of nondepressed motliers. Because we mini- Discussion mized the potential sociodemographic con- founds by having a homogeneous sample of As predicted, infants of depressed moth- depressed and nondepressed adolescent ers responded more positively to the rein- mofeeis of low SES, our results can be con- statement of touch following a still-face- sidered representative for ttiis particular no-touch episode than did infants of nonde- lower-income adolescent population. Given pressed motliers. Infants of depressed moth- this iiomogeneitj' of ova- sample, the results ers showed more positive affect (more may be limited in generalizability. smiles and vocalizations) and gazed more at their mothers' hands during the still-face- The effects observed in infant behavior with-touch period than the infants of nonde- were not accounted for by immediate group pressed mothers, who grimaced, cried, and differences in maternal behavior. That is, gazed away from their mothers' face more the depressed and nondêpressed mothers' often during this period. behaviors were not significantly different in We should note that, by specifically in- the conditions in which irifant behavior dif- troducing a still-face-with-touch period im- ferences \vere observed. This uniform pat- mediatelv- after a still-face-no-touch period, tern of maternal behavior was expected we were able to measure the soothing effects given the highly standardized procedures of of touch in the still-face situation right when this stiidy with bodi garoups of mothers. Both the infants began shovring the distressing ef- depressed and nonclepressed mothers were fects produced by their mothers' still-face speeiliqaliy instrBclgcI and given a demon- without touch. All infants became somewhat stratidrr showing tlieni how to behave in the similarly distressed during the still-face-no- still-fee-no-toueh period: and how to touch touch procedure; in particular, their gaze their infants during the still-face-with-touch away from mothers' face (gaze aversion) was period. In the abserice pf imrnediate ©roup significantly higher during this period com- differenees in matemai beha-vior, the differ- pared to the other three periods. When touch ences in the pattern of infant behaviors was inti-oduced in the still-face situation. across conditions can be related to the in- cc r-f CO -5^ too"co to o CS in in 00 to'r- S" >• CD !>. t-^ •v 35 00 i>- z 75 . oq22_Z! CO C^ §i OH OD S B 1

s CO -51" P" CO ^.-1 00 CtTtoP 00 00^vcin'i> to' res s 8. 9 (N '^ 00 f-H C35 CM »—1 t— "^ tC O CM f-H 00 I-H CO CD I>- o. oc I-H CSl ^"* O^l I-H I-H >-H CO oq_ m oq^ w • Q PB to -a OH S; a cq cq CO c: O CM "^ § i y; o o t^ CO '—1 o 05 cq in 35 to 00 o to 05 Csl — ^CO oq^co 2,"^ 2- tp z & 93 .

1 -a (X J3 i-H ,35 O! o"--! oq" .FACE-WIT I ^oqS? in S" o ©to C? > CO oq '—^ i>- O CO 35 ^CO C5 CO 10 CO CO CO -^ D. 35 c>q oq^

a Il l 1 O 1 35 CO <-0 m S'co £• CM CO

'i'A B & I-H 05 I—1

5; TIL L C/5 e an d 1

in ra c -g ^_^ ^_^ 1/: 01 ^^ ^~, S g c VI CO -^ 1—1 oc c CM t^ 00 c "H CO to CM rH CO (M CM 65 . Z (M cq —'00 ^ to CO ^'CO CO oc l 1 CO '—1 ~ 1—I r—' 1—1 r-H Q £1 T3 a; to : < X -a a 0 Jl oc CO CC CO "^ I— 1-H O o in ""^ CO ^'^ •—T a- —i O O ^ •^ t-- t-H , 1 , ; ^--tD CO ^ 1 I-H CO 00 i-H 1^ r"' cu CO "H 1—I C< C^ I— d CO oo ^ t^ C^l ^-'cM oq^ Q Q

3 1 s §

wl rt III o c > « ^ 5 E O £ ii g

MEA N IS. 1/ move m s e tou c ' s tota l V ' s voca l ' s face : •5 1, § C br, n 1 c .ir 1 d s tr Mot h z NOT E , 1790 Child Development fant's prior history of interactions with a de- who might normally be deprived of contin- pressed mother. The differences observed in gent maternal touch and contact at home. the infants' behavior may be attributed to Touch may have quickly become nonsooth- maternal depression and its concomitant his- ing, and perhaps aversive, for the infants of tory of interactions. nondepressed mothers, who might normally not be deprived and were more upset and Alternatives to learning-bistor>? explana- stressed by the preceding still-face-no-touch tions, however, should be considered. It has period. been argued that infents of depressed mothers are at unusually bigh risk for developing During the resumed normal play inter- depression due to genetic or prenatal. traBS- action, then, the infants of depressed moth- mission (Zuckerman, Ais, Baacbner, Parker, ers were not as distressed as those of nonde- & Cabrai, 1990). To predict infant "depressed mothers. Touch was initially pressive" behavioral outcomes from any sin- soo€iing to the infants of depressed mothers, gle factor, however, is almost impossible, and in the aftermath ofthe still-face periods, >vhether this factor is genetic or postnatal be- the infants of nondepressed mothers were bavior experience. To elucidate the etiology nidre apset. This phenomenon suggests that of infant depression was not the objective, as a result of their history of experiences, rather, our goal was to determine if a short the infaDts of depressed mothers were less intervention with touch by depressed moth- distressed by the still-fece perturbations and ers would increase 'infants' positive affect tiie absence of maternal toucli. Conceivably, and attention during still-face interactions. infents of depressed mothers could have As preclicted, touch was more effective in been less distressed in our study because enhancing the positive behavior in infants they received more optimal touch than they of depressed mothers. Depressed mothers \¥ere used to. seemed to lia^'c facilitated more positive af- The results ofthe present study can be fect and attention in their infants by toueh- related to findings from a recent learning ing tbem during -die iDteraetioiis, and tlie op- experiment using a synchronized reinforce- timal/nonintrusive type of touch used i3i this ment procedure (Pelaez-Nogueras et al., study appeared to provide comfort during 1996). Pelaez-Nogueras and colleagues the stressful still-faee interactions. found that contingent tactile stimulation by The findin,gs can be explained in a num- a caregiver during face-to-face interactions ber of waj'-s. One possible explanation is that increases affect and attention in .3-month-old the infants of nondepressed mothers did not infants. In that study, when touch was used show significantly less giiroaciiig and crying as part of the caregiver's social stimulation when touch was intioduced during the still and provided contingently, it effectively re- face because these infants wetele^s familiar inftirced and maintained higher rates of in- with maternal unavailabiliig-' (flat face and af- fant eye contact, smiles, and vocalizations. fect) and were thus much more difficult to InterestVagly, the infants in the present sootiie when touch was inti-oduced. More- study also increased eye contact with their over, infant grimacing and crj'ing continued moliers during the still-^faee-with-touch pe- to be emitted by the infants of nondepressed riod- tJnis may have occurred as a result of mothers even during the resinned nornial jntermittent contingent touch stimulation period. The increase in infant grimacing and on tafant making eye eonteict with their cr\'ing during the mothers' subsequent re- mothers. turn to noniial play following a period of ma- Bodi learning and emotional regulation ternal unavailability was also observed by processes prepare the infant to develop adap- Toda and Fogel (1993). tive and organized behavior strategies (Pel- Typically, mothers "fake good" and "try aez-5^ogueras, 1992; Thompson, 1994). The harder" to show positive behaviors during differences observed in the infants' behavior tlie initial moments of videotaping in experi- may lie in the different histories of mterac- ments. Thus, the absence of group differ- tions between mother and child and histories ences in maternal behavior in the initial of iufantbehavior regulation. The data for the normal play period should not be overinter- first normal play period show that infants of preted. For the purpose of this study, the depressed and nondepi'essed mothers dif- first normal play period may neither be rep- fered ill facial grimacing and looking at resentative of a "true" baseline nor as rele- hands, suggesting differences in their learn- vant as the subsequent differences obsewed ing Jiistories and conceivably in their ability later on tie final p;lay period after still-face to regulate their behaviors. However, even periods. It is possible that toucli was more though there were differences between in- soothing for infants of depressed mothers. faTits of depressed and nondepressed moth- Pelaez-Nogueras el al. 1791

ers, we should be cautious when attributing (New Directions for Child Development. No. these differences to the infants' prior interac- 34). San Francisco- Jossey-Bass. tive histories with their mothers in light ofthe Cohn, J. F., & Tronick, E. Z. (1983). Three-month- fact that there were no immediate group dif- old infants' reaction to simulated maternal de- ferences in the mothers' behaviors in the first pression. Child Development, 54, 185-193. normal play period. Cummings, E. M., & Cicchetti, D. (1990). Toward In sum, the effects of maternal touch a transactional model of relations between during still fece were more powerful for in- attachment and depression. In M. T. fants of depressed mothers than for infants Greenberg, D. Cichetti, & E. M. Cummmgs of nondepressed mothers, even when the (Eds.), Attachment in the preschool years. amount and type of touch provided by the Theory, research, and intervention (pp. 339- depressed and nondepressed mothers were 372). Chicago: University of Chicago Press. the same. Touch appears to have strong posi- Field, T. (1977). Effects of early separation, inter- tive influences on infant behavior, it can in- active deficits, and experimental manipula- crease positive affect, increase infants' nega- tions on infant mother face-to-face interac- tive affect, and direct infants' attention, in tion. Child Development, 48, 763-771. particular, the attention of infani5 of de- Field, T. (1984). Early interactions between in- pressed mothers during face-to-face interac- fants and their postpartum depressed moth- tions. The type of stimulation that involves ers. Infant Behavior and Development 7 touch during face-to-face interactions needs 527-532. to be investigated further. Although short- Field, T. (1986). Models for reactive and chronic term positive effects were achieved in the depression. In E. Tronick & T. Field (Eds.), present study, long-term assessments and Maternal depression and infant disturbance implementations of this type of intervention (pp. 47-60). San Francisco. Jossey-Bass. are needed to determine the more prolonged Field, T. (1992). Infants of depressed mothers. De- positive effects of touch on infant behavior. velopment and Psychopathology, 4, 49-66. Future research should focus on touch inter- Field, T., Healy, B., Goldstein, S.. & Guthertz, M vention strategies with infents and their de- (1990). Beha\dor state matching in mother- pressed mothers. infant mteractions of nondepressed versus depressed motJier-mfant dyads. Develop- References mental Psychology, 26, 7-14. Field, T., Schanberg, S. M., Scafidi. F, Bauer, Beardslee, W. R., Bemporad, J., Keller, M. B., & C R., Vega-Lahr, N., Garcia, R, Nystrom, J , Klerman, G. L. (1983). Children of parents & Kuhn, C. M, (1986). Tactile/kmesthetic with major depressive disorder: A review. stimulation effects on preterm neonates. Pedi- Amencan Journal of Psychiatry, 54, atrics, 77, 654-658. 1254--1268. Field, T , Vega-Lahr, N., Scafldi. F., & Goldstein, Beck, A. T., Ward, C. H., Mendelson, M., Mach, S. (1986). Effects of maternal unavailabilitv' J. E., & Erbaugh, J. (1961). An inventory for on mother-infant interactions. Infant Behav- measuring depression. Archive.s of General ior and Development, 9, 473-478. Psychiatry, i, 561-571. Gaensbauer, T J., Harmon, R. J., Cytryn, L, & Campbell, S. B., Cohn, J. F., & Meyers, T. (1995). McKnew, D. H. (1984). Social and affective Depression m first-time modiers: Mother- development in infant with manic-depressive infant interaction and depression chronicity. parent. American Journal of Psychiatry 141 Developmental Psychology, 31, 349-357. 223-229. Cohn, J. F., & Campbell, S. B. (1992, May). The Gelfand, D. M., & Teti, D M. (1990). The effects cour.se and correlates of depression in post- of maternal depression on children. Clinical partum women and their influence on infant Psychology Review, 10, 329-353. socio-emotional development. Symposium Gewirtz, J. L., & Pelaez-Nogueras, M. (1992). In- delivered at the International Conference on fant social referencing as a learned process. Infant Studies, Miami. In S. Feinman (Ed), Social referencing and Cohn, J. F., Campbell, S. B , Matias, R., & Hop- the social construction of reality in infancy kins, J (199QJ. Face-to-face interactions of (pp. 151-173). New York. Plenum. postpartum and nondepressed motlier-mfant Goodman, S (1992). Understanding the effects of pairs at tvpo months. Developmental P.sychol- depressed mothers on their children. In E. F. ogy, 26, 18.5-193. Walker, B. Comblatt, & R. Dworkin (Eds ), Cohn, J. F., Matias, R., Tronick, E. Z., Connell, Progress in experimental personality and D., & Lyons-Ruth, K. (1986). Face-to-face in- psychopathology research (Vol. 15, pp 47- teractions of depressed mothers and their in- 109). New York: Springer. fants. In E. Z. Tronick & J. Field (Eds.), Ma- Gnsella, J. L., Muir, D. W., & Tronick, E. Z. terrtal depression and infant disturbance (1988) The effect of manipulating maternal 1792 Chad Development behavior during an interaction on 3- and 6- fant Behavior and Development, 13, month-olds' affect and attention. Child Devel- 167-188. opment, 59,1111-1124. Stack, D. M , & Muir, D. W. (1990). Tactile stimu- Guthertz, M., & Field, T. (1989). Lap computer or lation as a component of social interchange. on-line coding and data analysis for laboratory New interpretations for the still-face effect and field observations. Infant Behavior and British Journal of Developmental Psychol- Development, li, 305-319. ogy, S, 131-145. Hammen, C. (1992). Cognitive, life strtess, and in- Steck, D. M., & Muir, D. W. (1992). Adult tactile terpersonal a.pproaches to a developmental stimulation during face-to-face interaction psycliopatholog>' model of depression. Devel- modulates five-month-olds" affect and atten- opmertt and Psychopathology, 4, 189—206. tion. Child Development, 63, 1509-^1525. Lamb, M. E., Morrison, D. C, & Malkin, C. M. Teti, D. M., Gelfand, D. M., Messinger, D. S., & (1987). The development of infant social ex- Isabella, R. (1995). Maternal depression and pectations in face-to-faoe interaction: A longi- the quality of early attachment: An exaniina- tudinal study. MerriR-Pahner Quarterlij, 33, tion of infantSj, preschoolers, and their moth- 24i-2S4. ers. Oetjelopmental Psychology, 31, 1—13. Lyons-laut:h, K., Zoll, D., Connell, D., & Gmne- Thompson, R. (1994). EiHOtion regulation: A baum, H. U. (1986). The depressed mother tlieme in search of definition. In N. A. Fox and her one-year-old infant: Enviromnent, in- (Ed.), The development: of emotion regula- teraction, attachment, and infant develop- tion: BialogicEil and behavioral consider- meJHt. in E. Z. Tronick &T. Field (Eds.), Ua- ations. Monographs of the Sodeiy far Re- terrml depression and infant distitrbance search in Child Development, 59(2-3, Serial (New Directions for Child Development, No. PJo. 240). 34). San Francisco: Jossey-Bass. Toda, S., & Fogel, A. (1993). Infant response to Malphurs, J., Raag, T., Field, T., Pickens, J., & the still-face situation at 3 and 6 months. De- Pekez-Nogueras, M. (1996). Touch by intru- velopmental Psychology, 29, 532-^38. sive and withdrawn niiothers with depressive Tronick, E. 2., Ais, H., Adamson, L., Wise, S., & symptoms. Early Development and Parent- lirazelton, T. B. (1977). The infant's response ing, a. 122.1-5. to entrapment between contradictory mes- Mayes, L. C, & Carter, A. S. (1990). Emerging sages in face-to-face interaction. Journal of social regulatory capacities as seen in the ChUd Psychiatry, 1", 1-13. still-face situation. Child Development, 61, Tronick, E. Z., & Gianino, A. F. (1986). The trans- 754-t63. mission of maternal disturbance to the infant. Ottenbacher, K. J., Muller, L., Brandt, D., Heint- In E. Z. Tronick & T. Field (Eds.), Maternal zelman. A., Hojem, P., & Sharpe, P. (1987). depression and infant disturbance (New Di- Tfie efFectiveness of tactile stimulation as a rections for Child Development, No. 34). San form of early intervention: A quantitative Francisco. Jossey-Bass. evaluation. Developmental and Behavioral Weissriian, M. M., Prusoff, B. A., Gammon, G. D., Pediatrics, 8, 68-76. Merikangas, K. R., Leckman, J. F., & Kidd, Pelaez-Kogueras, M. (1992). Infant learning to K. K. (1984). Psychopathy in children (ages reference maternal emotional cues. Unpub- 6-18) of depressed and nornial parents. Jour- lished doctoral dissertation, Florida Interna- nal of American Amdemu of Child Psychia- tional University. tnj, 23, 78-84. Pelaez-Nogueras, M., Field, T., Cigales, M., Gon- Whiffea, V. E., & Gottlib, 1. M.;( 1989). Infants of palez. A., & Clasky, S. (1994). Infants of de- postpartum depressed mothers: Tempera- pressed mothers show less "depressed" be- ment and cognitive status, journal of Abnor- havior with their nursery teachers. Infant mal Paychoiogif, 98, 274-279. Uental HeaUkJournal, 15, 358-367. Winer, B.J. (1971). Statistical principles in exper- Pelaez-Nogueras, M.,