Depressed Mothers' Touching IncreasesInfants' Positive Affect and Attention inStill-Face Interactions

Martha Pelaez-NoguerasFlorida International University

Tiffany M. FieldUniversity of Miami

Ziarat HossainFort Lewis College

Jeffrey PickensJames Madison University

PELAEZ-NoGUERAS,MARTHA;FIELD,TIFFANYM.; HOSSAIN,ZIARAT;and PICKENS,JEFFREY.De-pressed Mothers' Touching Increases Infants' Positive Affect and Attention in Still-Face Interac-tions. CHILDDEVELOPMENT,1996, 67, 1780-1792. The effects of depressed mothers' touchingon their infants' behavior were investigated during the still-face situation. 48 depressed andnondepressed mothers and their 3-month-old infants were randomly assigned to control andexperimental 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 nondepressedmothers were instructed and shown how to provide touch for their infants during the still-face-with-touch period. Different affective and attentive responses of the infants of depressed versusthe infants of nondepressed mothers were observed. Infants of depressed mothers showed morepositive affect (smiles and vocalizations) and gazed more at their mothers' hands during thestill-face-with-touch period than the infants of nondepressed mothers, who grimaced, cried, andgazed away from their mothers' faces more often. The results suggest that by providing touchstimulation for their infants, the depressed mothers can increase infant positive affect and atten-tion and, in this way, compensate for negative effects often resulting from their typical lack ofaffectivity (Hat facial and vocal expressions) during interactions.

Early interaction disturbances place in-fants of depressed mothers at risk for lateraffective and socioemotional disorders(Field, 1992; Gaensbauer, Harmon, Cytryn,& McKnew, 1984; Zahn-Waxler, Cummings,McKnew, & Radke-Yarrow, 1984). Having adepressed mother increases by three times achild's risk of developing the abnormalitiescharacteristic of depressed mothers (Weiss-man et aI., 1984). Numerous studies havedocumented the negative impact of mater-

nal depression on early infant interactionsand development and have identified thebehavior patterns of depressed mothers asunresponsive, insensitive, ineffective, non-contingent, emotionally flat, negative, disen-gaged, intrusive, avoidant of confrontation,and generally less competent and unin-volved with their infants (e.g., Campbell,Cohn, & Meyers, 1995; Cohn, Matias, Tron-ick, Connell, & Lyons-Ruth, 1986; Cohn &Tronick, 1983; Field, 1984, 1986; Lyons-

The authors wish to thank Julie Malphurs, Jeanette Gonzalez, Claudia Larrain, and AngieGonzalez, for aiding with data collection and Roberto Pelaez for help with data analyses. Theresearch reported in this article was supported by National Institute of Mental Health ResearchScientist Award MH00331 and National Institute of Mental Health Basic Research GrantMH46586. Portions of these data were presented in March 1993 at the meeting of the Society forResearch in Child Development, New Orleans, Louisiana. Correspondence and reprint requestsshould be addressed to Dr. Martha Pelaez-Nogueras, Department of Educational Psychologyand Special Education, Florida International University, Miami, Fl 33199.

[Child Development, 1996,67,1780-1792. © 1996 by the Society for Research in Child Development, Inc.All rights reserved. 0009·3920/96/6704·0031$01.00]

Ruth, Zoll, Connell, & Grunebaum, 1986;Pelaez-Nogueras, Field, Cigales, Gonzalez,& Clasky, 1994).

Infants of depressed mothers, in turn,appear to develop a depressed mood style asearly as 3 months. The "depressed" infantstypically exhibit less attentiveness, fewersmiles, more fussiness, more gazing away,and lower activity levels when interactingwith their depressed mothers than infants ofnondepressed mothers (Cohn, Campbell,Matias, & Hopkins, 1990; Gelfand & Teti,1990; Goodman, 1992). Moreover, maternaldepression has been significantly associatedwith attachment insecurity among infantsand preschoolers (Teti, Gelfand, Messinger,& Isabella, 1995). Infants of depressed moth-ers, however, do not necessarily generalizetheir "depressed mood" to other adults.When the infants of depressed mothers in-teracted with their nondepressed nurseryteachers, the infants' behavior recovered,and their activity levels and positive affectrates were higher than when interactingwith their depressed mothers (Pelaez-Nogueras et aI., 1994).

In general, depressed mothers and theirinfants appear to share their behavior states,spending more time in negative attentive/affective behavior states than nondepressedmother-infant dyads (Field, Healy, Gold-stein, & Guthertz, 1990). Different profilesof behavior have been identified, includingdisengaged mothers (withdrawn and pas-sive) and intrusive mothers (e.g., angry facialexpressions and intrusive poking of the in-fant) (Field et aI., 1990; Malphurs, Raag,Field, Pickens, & Pelaez-Nogueras, 1996).But despite the variability observed in themothers' interaction styles, the infants ofboth disengaged and intrusive mothers areusually uniformly distressed. Also, whetherexperiencing postpartum or chronic depres-sion (Campbell et aI., 1995), the commonfinding in the literature is that the depressedmothers' negative mood states and lack ofaffective responses negatively affect thechild's behavior. In this way, the infants ofdepressed mothers begin to show growthand developmental delays at 1 year if theirmothers remain depressed over the firstyear. Normally, the developmental delaysare manifested by inferior performance onBayley Mental and Motor Scales at 1 year ofage, but other behavioral deficits have alsobeen noted, including heightened emotion-ality and a lower level of symbolic play(Field, 1984;Gaensbauer et aI., 1984; Samer-off& Seifer, 1983;Whiffen & Gottlib, 1989).

Researchers have prospectively studiedinfants and toddlers of depressed mothers toanalyze the processes and mechanismswhereby depression may affect infant be-havior. Diverse mechanisms have been hy-pothesized to produce the negative out-comes observed in infants and children ofdepressed mothers (e.g., Beardslee, Bemp-orad, Keller, & Klerman, 1983; Cummings &Cicchetti, 1990; Hammen, 1992). However,elucidation of the mechanisms and pro-cesses involved in the transmission of soci-oemotional behavioral problems from de-pressed mothers to their infants is still amajor challenge for developmental research-ers. This is because early development ofinfant depression may result from the inter-action of multiple influences, including bio-logical factors and psychosocial factors. Sev-eral models of early development ofdepression have been proposed, includingmutual regulation (Tronick & Gianino,1986), multivariate cumulative risk (Field,1992), and temperament and genetic predis-position (Whiffen & Gottlib, 1989). Thesemodels have focused on the effects of multi-ple factors that include prenatal influencesand postnatal experiences. An infant show-ing a depressed-mood pattern could be(a) biologically predisposed to depressiondue to prenatal exposure to the depressedmother's physiological imbalance and hor-monal status or due to a potentially congeni-tal predisposition or (b) environmentally af-fected due to continuing maternal depressedbehavior patterns, like unresponsivenessand flat affect. Thus, multiple factors seemto be affecting both mothers' depression andinfants' behavioral patterns denoting "de-pression," and there are multiple interac-tions between these underlying aflective,perceptual, physiological, and biochemicalprocesses.

The objective of the present study wasto determine whether depressed motherscan improve their infants' attentive and af-fective responses by providing touch stimu-lation during still-bce interactions. Touch,as a source of stimulation, has received littleattention in the mother-infant interaction lit-erature. The studies reported below suggestthat tactile stimulation is a significant con-tributor to infant growth and social devel-opment.

Touch Stimulationin High Risk Populations

Studies involving touch of premature in-fants and neonates have reported improve-ments in physiological growth, motorIre flex,

cognitivellanguage, and visual/auditory de-velopment (see Ottenbacher et aI., 1987).For example, an intervention program ofstroking and passive limb movements threetimes daily for a total of 45 min per day ledto increased weight gain, increased wake-fulness and activity level, and improved per-formance of pre term infants on the Brazeltonscales of orientation, motor, and range ofstate behavior (Field, Schanberg, et aI.,1986). A follow-up study suggested continu-ing advantages for the stimulated infants, in-cluding better growth and development(Scafidi et aI., 1990).

Touch during Face-to-Face InteractionsFace-to-face interactions are a primary

way behavior disorders seem to be transmit-ted from mother to infant (Cohn et aI., 1986;Field, Vega-Lahr, Scafidi, & Goldstein,1986). The quality of infant behavior hasbeen related to the unresponsiveness andemotional unavailability of their mothersduring these interactions (Sameroff & Seifer,1983; Tronick & Gianino, 1986). Only a fewstudies have investigated the effects of ma-ternal touch during face-to-face interactionswith their infants or the use of touch inter-ventions to facilitate better interactions be-tween mothers and their infants (e.g., Field,1977; Malphurs et aI., 1996; Pelaez-No-gueras et aI., 1996; Stack & Muir, 1990,1992).

For example, Pelaez-Nogueras et aI. (inpress) found that touch can reinforce andmaintain high rates of infant eye contact re-sponses, vocalizations, and smiles duringface-to-face interactions. In that study, usinga synchronous-reinforcement operant proce-dure, touch stimulation (gentle rubbing ofthe infant's arms, legs, and feet) was pro-vided by a caregiver while the infant wasmaking eye contact with her. After severalconditioning sessions, the infants showedpreferences for the reinforcing stimulationthat included touch, as shown by the factthat they smiled and vocalized more andmade more eye contact with the caregiver.Those findings suggest that infants' attentionand positive affect can be reinforced andmaintained by an adult providing contingenttactile stimulation during face-to-face inter-actions.

Infants' affective responses to stressfulevents like the still face of their mothersduring interactions have also been investi-gated (Cohn & Tronick, 1983; Tronick, Als,Adamson, Wise, & Brazelton, 1977). In thestill-face procedure, the mother's behavior ismanipulated by having her adopt a station-

ary, expressionless poise. The still-face pro-cedure has been used to study mother-infantinteractions and to examine the effects ofmaternal behavior on infant affect and atten-tion (e.g., Gusella, Muir, & Tronick, 1988;Lamb, Morrison, & Malkin, 1987; Mayes &Carter, 1990; Toda & Fogel, 1993) and oninfant social referencing (Gewirtz & Pelaez-Nogueras, 1992). During the still-face situa-tion the continuation of maternal gaze to-ward the infant, coupled with her lack ofresponding and lack of touch, may lead to aninfant reacting with negative affect and othercoping behaviors. Stack and Muir (1990)found that when mothers were asked to befacially unresponsive, silent, and not totouch their infants during the still-face epi-sode, infants displayed more grimacing andless smiling compared to periods of normalinteraction. However, when touch was intro-duced during the still-face period, infants'positive affect and attention was higher. Itremains to be determined, however,whether infants of depressed mothers aremore sensitive to maternal touch than in-fants of nondepressed mothers. As yet, nostudies have investigated the effects of touchby depressed mothers using the still-faceprocedure.

The present study was designed to testthe hypothesis that depressed mothers canreduce the negative effects elicited by theirstill faces by providing additional touch fortheir infants. We thought it important to ex-amine if touch provided by depressed moth-ers can help their infants to regulate affectbehavior and increase their attention. Therationale was that infants of depressed moth-ers would not be as distressed as the infantsof nondepressed mothers during the still-face-with-touch situation because they wereused to seeing their mothers with flat affect.We expected that for these infants, touchcould minimize (or compensate for) the lackof stimulation from the other sources (i.e.,voice 'llnd face). On the other hand, the in-fants of nondepressed mothers were ex-pected to be more difficult to soothe, evenafter touch was introduced in a still-face-with-touch period, because their mothers'continuous still face was so unexpected andatypical in their experience.

The main assumption was that, for thedepressed group only, even when motherscontinue displaying flat affect, their use oftouch could decrease the negative effects ofthe still-face condition. Due to their historyof repeated exposure to unresponsive mater-nal behavior, it was thought that infants ofdepressed mothers would respond more

positively than infants of nondepressedmothers when optimal touch (mild strokes/movelJlents) was introduced in the still-facesituation. Thus, maternal behavior was ma-nipulated, touch was standardized, and themain grouping (independent) variable wasmaternal depression score. By standardizingtouch we minimized differences in de-pressed and nondepressed mothers' kinds oftouch and were able to assess whether ma-ternal depression could account for the dif-ferences in infant behavior.

MethodSubjects

Forty-eight 3-month-old infants (meanage = 13.5weeks, SD = 1.2) and their moth-ers (mean age = 19.1, SD = 2.7) partici-pated in this study. All infants were healthy,born at gestational age (M = 38 weeks),were of normal birthweight, and had no his-tory ofmedical complications. Subjects wererecruited from a longitudinal study sampleof low socioeconomic status based on thetwo-factor Hollingshead Index. Motherswere primiparous black (53%), Hispanic(40%), or Caucasian (7%), and were single(86%) adolescents, and their infants werenormal full term infants. Three mother-infant dyads needed to be rescheduled be-cause the babies were fussy and sleepy.

Mother-infant dyads were assigned toone of four groups: Depressed Mothers-Experimental (N 16), NondepressedMothers-Experimental (N = 16), and De-pressed Control and Nondepressed Control(N = 16). The Beck Depression Inventory(BDI) scores defined the depressed and thenondepressed groups (depression classifica-tion is described in detail in the followingsection). The groups did not differ on demo-graphic variables, including age, ethnicity,marital status, and SES, resulting in a homo-geneous sample. To ensure group equiva-lence, infants and their mothers were as-signed to control or experimental groupsthrough a random stratification procedure,stratifying in accordance to maternal depres-sion score.

ProcedureMaternal depression assessment.-

Assignment to the depressed groups wasbased on cutoff scores on the Beck Depres-sion Inventory (BDI; Beck, Ward, Mendel-son, Mach, & Erbaugh, 1961). The 21 BDIitems are scored on a four-point scale indi-cating absence/presence and severity of de-pressed feelings, behaviors, and symptoms.The scale is among the commonly employed

instruments in research on nonclinically de-pressed samples. This self-report scale wasused rather than a diagnostic interview be-cause Cohn and Campbell (1992) have re-ported that depressed mothers' interactionbehaviors are more highly correlated withself-report depression scores than they arewith diagnostic interview measures. Moth-ers with BDI scores of 13 or greater (cutpointof depression in most research protocols)were assigned to the depressed group andmothers with scores of 9 or less were as-signed to the nondepressed group. We ad-ministered the BDI to 61 mothers to yieldour sample of 24 depressed mothers. In pre-vious studies with this population, approxi-mately 30% of the mothers sampled receivedscores greater than 16 on the BDI (e.g., Fieldet aI., 1990). The mean BDI score for all de-pressed mothers in our sample was 21 (SD= 9.1), ranged from 13 to 52, and for thenondepressed mothers was 4.1 (SD = 2.7),ranged from 1 to 9. Mothers with BDI scoresof zero, 10, 11, and 12 did not participate inthis study. The BDIs were administered 15min before the interaction in a waiting roomnext to the laboratory by a research assistant.

Apparatus and setting.-Infants wereseated in an infant seat facing their mothersat a distance of approximately 15 inches.Mothers were seated directly facing their in-fants at eye level. Two cameras, located oneither side of the mother-infant dyad, wereconnected to a video recorder and a specialeffects generator to yield a split-screen im-age. One camera recorded the frontal viewof the infant, and the second camera re-corded the mother's face and hands. A time-date generator connected to the monitor wasused to time the duration (in minutes, sec-onds, and milliseconds) of each period forsubsequent coding.

Design.-A repeated-measures be-tween-groups design was implemented: twogroups (depressed vs. nondepressed) x twoconditions (control vs. experimental) x foursuccessive periods: (A) 90-sec normal inter-action, followed by (B) 90-sec still-face-no-touch, (C) 90-sec still-face-with-touch, andfinally, (A) 90-sec normal interaction. Six-teen additional mother-infant dyads wereused as a no-still-face control group. Mothersin the control condition only received thenormal interaction instructions across thefour consecutive periods of the study. Thedesign compares the 16 controls (half de-pressed and half nondepressed) to the 32 ex-perimental (half depressed and half nonde-pressed). The order of the periods was not

counterbalanced because the purpose was tohave a still-face-no-touch period precedinga still-face-with-touch period specifically toinduce distress in order to increase thechances of getting an effect and thus to com-pare the depressed and nondepressed dy-ads' performances.

Instructions.- The total procedure re-quired approximately 8 min. During the four30-sec intervals between the four periods allmothers were given instructions. Instruc-tions were standard for all mothers. To ad-dress the question of whether infants of de-pressed and nondepressed mothers responddifferentially to touch when their motherspose a still face, it was important to reducevariability in mothers' behavior during thestill-face situation. For this reason, we im-posed still-face instructions to all mothers inthe experimental condition.

Before the first normal play period of in-teraction, mothers in the experimental con-dition were instructed to play with their in-fants as they would normally do at home.For the second period (still-face-no-touch),these mothers were instructed to look/gazeat their infants with a neutral expression,and to refrain from speaking, smiling, andtouching the infant during this period. Forthe third period of interaction (still-face-with-touch) instructions were given to look/gaze at the infant with a neutral expression,to refrain from speaking and smiling, but totouch the infant as modeled. In the last nor-mal period, mothers received the same in-structions as in the first normal period.

To ensure that mothers maintained astill face throughout the still-face periods,continuous monitoring was conducted bythe second research assistant observing theinteraction from the observation room. Theobserver constantly checked that motherswere complying with instructions and werenot making any change in facial expressions,thus, maintaining their "neutral" still facewhile gazing at the infant. In addition, wemonitored that mothers were not cooing orvocalizing and that their touch was the sameas instructed. All mothers in both groupscomplied with instructions (> 90% of thetime). In those cases where mothers werenot following instructions and smiled, vocal-ized, or touched their infants incorrectly, thesession was interrupted and postponed for asecond visit when further training was pro-vided. If an infant was showing signs of be-ing distressed during any of the four experi-mental periods or cried consistently for more

than 15 see, the session was interrupted andrescheduled. A total of five mother-infant dy-ads needed to be retrained and rescheduledfor a second visit.

Touch procedure.-Just before the still-face-with-touch period, all mothers in theexperimental condition received a briefdemonstration of optimal touch. The "opti-mal" touch procedure involved a motherstroking and rubbing rhythmically the in-fants' arms, legs, and feet using the five fin-gers of both hands for the duration of thestill-face period (90 see). The experimentermodeled gentle pressure in slow circularmotions at a rate of approximately one circu-lar rub per sec. Negative touch was avoided.Negative touch involves rough tickling, pok-ing and tugging while interacting with theinfant, including poking the baby's face,arms, or stomach, or pinching or squeezingthe infant, or pulling or shaking the infant.Mothers were instructed not to tickle orpoke their infants during this procedure, norto pull intensively their infants' legs or arms.The mothers' touch was checked routinelyduring the interactions to make sure theywere providing touch as instructed.

Behavior coding.- The onset and offsetof the videotaped behavior were registeredby pressing numeric codes on a laptop com-puter. All behavior modalities were codedseparately. The behaviors were coded con-tinuously and featured a second-by-secondlisting of behaviors and a matrix of percent-age time the behaviors occurred (Guthertz& Field, 1989).One view of the videorecordwas used per each modality: (1) infant facialexpressions (three codes: smile, neutral, gri-mace), (2) infant vocal expressions (threecodes: positive vocalizations, no vocaliza-tions, and protest/crying), (3) infant gaze be-havior (used three codes: gaze at mothers'face, gaze away from mother's face, gaze atmother's hands). Thus, coding of these mea-sures required three separate viewings ofeach record. In this way, coding of the in-fants' behaviors included three positive be-haviors: (1) smiling, (2)vocalizing, (3)gazingat mothers' hands, and three negative af-fective behaviors: (4) crying, (5) grimacing,and (6) gazing away from mom.

For infant smiling to be coded the infantmouth had to be "upturned," whether themouth was open or closed. For infant gri-macing, the infant's mouth had to be turneddown or curled or the infant had to be cry-ing. For gazing away from the mother, theinfant had to be looking at any other place

but the mother's face, hands, or body. Posi-tive vocalizations were discrete sounds likethose involved in cooing and babbling (butthe infant could not be fussing or protesting).For crying, the infant had to be grimacingand emitting nondiscretelloud sounds.

Given the highly standardized proce-dure of this study, for control purposes themother's behaviors were also coded; that is,mother's touch, facial expressions (smiles,negative/angry, neutral), and mother's vocalsounds were coded. This allowed us to en-sure that mothers were following the still-face, no voice, and no-touch and touch in-structions. For touch behavior, five(numeric) codes on the laptop computerwere used to code touch behavior: (1)mother's hand resting on baby, (2) mildtouching (stroking, caressing, rubbing), (3)intense touching (tickling, poking), (4) mildmovement (lifting baby's feet or arms inslow, rhythmic cycling), and (5) intensemovement (quick intense movements ofarms and legs or pulling arms or legs) (Stack& Muir, 1990). The purpose to measuretouch was to ensure that mothers were pro-viding mild touch and movements (2 and 4)for at least 75% of the time during the still-face touch period and were not making in-tense movements or pulling the infant's legsor arms.

Observer ReliabilityObservers were unaware of the hypoth-

eses and of the mothers' depression status.The two independent raters were trained to90% reliability on each response categorywith an experienced rater. Reliability of thebehavior measures was determined on one-third of the sample. Product-moment corre-lation coefficients were obtained on the per-centage scores of primary and secondaryobservers on all response measures of infantand mother behaviors. Observer reliability,calculated separately for each response mea-sure, was at p < .001 for each measure. Thereliability coefficients obtained for infants'behaviors were as follows: infant smile, r =.96; infant vocalization, r = .92; infant gazeat hands, r = .90; infant grimacing, r = .97;infant crying, r = .94; infant gaze away,r = .98. For mothers' behaviors the reliabil-ity coefficients were: (1) mother's hand rest-ing on baby, r = .99; (2) mild touching, r =.95; (3) intense touching, r = .92; (4) mildmovement, r = .96; and (5) intense move-ment, r = .93; vocal sounds, r = .98; smiles,r = .92; negative/angry face, r = 88; andneutral face, r = .96.

Results

The first analyses were a 2 (group: de-pressed vs. nondepressed) x 2 (condition:experimental vs. control) x 4 (periods of in-teraction) MANOVAs on infants' positive be-haviors (smiling, vocalizations, and gazing atmothers' hands), and on infant's negative be-haviors (grimacing, crying, and gaze awayfrom mothers). For the first MANOVAon in-fant positive behaviors, the analyses yieldeda significant three-way interaction effect ofgroup x condition x periods, F(9, 36) =2.56, p < .05. Then, significant main effectswere also observed for group, F(3, 42) =4.28, p < .01, and condition, F(3, 42) = 2.7,p = .05. For the MANOVA on negative be-haviors, the analyses yielded a significantthree-way interaction effect of group x con-dition x periods, F(9, 36) = 2.56, p < .05.For the negative infant behaviors significantmain effects were also observed for group,F(3, 42) = 3.90, p < .05, and for condition,F(3, 42) = 2.68, p = .05.

Separate analyses for the control and theexperimental conditions revealed: (1) nochanges in the control condition on any be-havior were observed over time, across thefour periods; (2) no significant differences inthe behavior of the infants of depressed andnondepressed mothers were observed in thecontrol condition across the four periods;(3) no differences were observed betweenthe control and experimental mother-infantdyads in the first normal period. These anal-yses suggested that the control and experi-mental conditions were similar at the begin-ning (first normal period) of the study andthat the infants were not fatigued over time.

A significant main effect of group (de-pressed vs. nondepressed), F(3, 28) = 7.42,p

peated measures showed significant interac-tion effects (Winer, 1971). Paired t tests werealso conducted within subjects to comparethe between still-face-with-touch and still-face-no-touch periods (shown by subscriptsin Table 1). The results on each variablefollow.

Infant Positive BehaviorsSmiling.-ANOVA results for smiling

yielded a group trend (depressed vs. nonde-pressed), F(l, 30) = 3.68, p < .06, and a sig-nificant group x period interaction effect,F(3, 90) = 2.94, p < .05. The proportion ofsmiling decreased from the normal period ofinteraction to the still-face-no-touch periodin both groups (Table 1). However, only thedepressed group showed a significant in-crease in smiling from the still-face-no-touchto the immediately following still-face-with-touch period, t(15) = - 2.33, p < .05. Simplemain effect tests performed on infant smilingrevealed that the depressed and nonde-pressed groups differed in the still-face-with-touch period, F(l, 30) = 11.15, p <.005, and in the last normal period, F(l, 30)= 16.00, p < .005, with the depressed groupsmiling more.

Vocalizations.-An ANOVA yielded asignificant group effect (depressed vs. non-depressed), F(l, 30) = 4.79, p < .05, and asignificant group x period interaction effect,F(3,90) = 8.74, p < .001. The proportion oftime spent vocalizing decreased from thefirst normal period of interaction to the still-face-without-touch period for the nonde-pressed group only (Table 1). Post hoc sim-ple main effects revealed group differencesin vocalizing during the still-face-with-touchperiod, F(l, 30) = 5.52, p < .05, and duringthe last normal period, F(l, 30) = 12.50,p < .001, with infants of depressed mothersvocalizing more than infants of nonde-pressed mothers. The differences noted ininfants' vocalizations in the normal periodswere not significant.

Gazing at mother's hands.-ResuIts re-vealed a main effect for group (depressed vs.nondepressed), F(l, 30) = 10.90, p < .005,and a group x period interaction effect, F(3,90) = 3.72, p < .01, in gaze at mother'shands. As expected, the proportion of timeinfants gazed at mother's hands increasedsignificantly from the still-face-no-touch pe-riod to the still-face-with-touch period forboth the depressed group, t(15) = -4.88,p < .001, and for the nondepressed group,t(15) = 5.53, p < .001. Simple main effectsanalysis revealed that infants of depressedmothers gazed at their mothers' hands more

often than infants of nondepressed mothersduring the first normal episode, F(l, 30) =5.10, p < .05, still-face-with-touch period,F(l, 30) = 4.97, p < .05, and during the lastnormal period, F(l, 30) = 13.30, p < .001.Infant Negative Behaviors

Grimacing.-An ANOVA on grimacingyielded a group effect (depressed vs. nonde-pressed), F(l, 30) = 5.50, p < .05, and agroup X period interaction effect, F(3, 90)= 7.01, p < .001. For the depressed group,infant grimacing decreased from the still-face-no-touch period to the still-face-with-touch period in the depressed group only,t(15) = 2.58, p < .05. Simple main effectsanalysis revealed that infants in the de-pressed group grimaced less often than theinfants of nondepressed mothers during thestill-face-with-touch period, F(l, 30) =11.15, p < .005. Grimacing was also less fre-quent, F(l, 30) = 7.62, p < .01, in the de-pressed group compared to the nonde-pressed group during the last normal period.

Crying.-For crying, only a group x pe-riod interaction effect was obtained, F(3, 90)= 2.92, p < .05. For the depressed grouponly, infant crying decreased from the still-face-no-touch period to the still-face-with-touch period, t(15) = 3.43, p < .005. Cryingwas lower, F(l, 30) = 4.98, p < .05, in thedepressed group compared to the nonde-pressed group during the still-face-with-touch period. Crying continued to be lower,F(l, 30) = 4.39, p < .05, for the depressedgroup compared to the nondepressed groupduring the last normal period.

Gazing away from mother.-AnANOVA yielded a group effect (depressedvs. nondepressed), F(l, 30) = 5.79, p < .05,and a group x period interaction effect, F(l,30) = 7.55, p < .001. The proportion of timethe infants gazed away from their motherssignificantly decreased, t(15) = 5.65, p <.001, from the still-face-no-touch to the im-mediately following still-face-with-touch pe-riod in the depressed group but not in thenondepressed group. Simple main effectsanalysis performed on gazing away from themother revealed that the nondepressedgroup gazed away more than the depressedgroup during the still-face-with-touch pe-riod, F(l, 30) = 18.00, p < .001, and alsoduring the last normal period, F(l, 30) =19.14, p < .005. The difference observed inthe first normal period was not significant.

Mothers' BehaviorRepeated-measures MANOVA on ma-

ternal touch revealed no significant main ef-fects of groups (depressed vs. nondepressed)

TABLE

1

MEA

NPER

CEN

TAG

EO

FI""FA

NTS'(N

=16

Depressed

andN=

16N

ondepressed)BEH

AV

IORS

ACRO

SSFO

UR

SUCCESSIV

ECO

ND

ITION

S

NO

RM

AL

PLAY

STILLFA

CE-N

o-ToUCH

STILL-FA

CE-W

ITH-TO

UCH

NO

RM

AL

PLAY

INFA

NTS

Non-

Non-

Non-

Non-

EFFECT

BEH

AV

IORS

Depressed

depressedD

epresseddepressed

Depressed

depressedD

epresseddepressed

p

Smile

..............29.0_

33.7.7.3_

2.6b

18.6_2.8

b32.1_

16.7cI*

(24.6)(24.9)

(10.6)(3.3)

(18.6)(3.5)

(25.0)(17.3)

Vocalizations

........10.5_

20.5_1O

.9_b

6.5b

17.3_c4.8

b27.1c

4.1b

G*I****

(9.1)(22.0)

(15.0)(6.3)

(20.7)(4.8)

(22.6)(4.5)

Grim

ace............

5.0_.O

b5.6_

11.6_1.3

b8.1_c

2.5_b

32.5d

G*I****

(9.3)(.0)

(8.6)(19.1)

(1.9)(10.5)

(4.1)(43.3)

Crying

.............1.6_

.2.8.2

b1O

.6b

1.1_1O

.1b

2.9_13.4

b1*

(2.9)(.5)

(10.2)(16.2)

(1.6)(16.2)

(4.9)(19.3)

Gaze

athands

.......15.3_

4.8b

1.8c.4c

38.5b

19.6_25.5_

2.4b

G***I**

(20.3)(6.4)

(3.0)(.6)

(30.8)(14.0)

(25.1)(3.3)

Gaze

away

..........14.3_

10.3_55.4

b53.3

b17.8_

46.4b

14.2_39.3

bG

*I****(15.9)

(8.3)(30.9)

(30.9)(13.2)

(23.5)(13.1)

(18.8)

NO

TE.-M

eansbearing

differentsubscripts

aredifferent

atp<

.05or

lessrevealed

bypost

hoccom

parisonof

adjacentgroups.

G=

groupeffect.

I=

group(depressed/nondepressed)

xperiod

interactioneffect.

Standarddeviations

arein

parentheses.*p<

.05.**

p<

.01.***

p<

.005.****

p<

.001.

or group X period interaction effects (p >.10). Touch was provided almost continu-ously by all mothers (97.7% of the time bydepressed mothers and 93.7% of the time bynondepressed mothers) during the 90-secstill-face-with-touch period (Table 2). Theinstructions provided to both groups for thestill-face-no-touch and for the still-face-with-touch periods minimized any potentialdifference in maternal behavior. Also, touchinstructions seemed to produce a carryovereffect of the mothers' touch behavior fromthe still-face-with-touch period to the lastnormal period of interaction, during whichboth depressed and nondepressed motherstouched their infants more than during thefirst normal play period. The higher amountof touch during the last normal period com-pared to the first normal period could haveaccounted for the significant differences ob-served in infant behaviors between thesetwo periods.

Overall MANOVA for mothers' vocalsounds, smiles, and stillface revealed no sig-nificant main effects of groups (depressed vs.nondepressed) or group X period interac-tion effects (p > .10). This result was alsoexpected given that the maternal behaviorin both groups was highly standardized andunder experimental control during the still-face-no-touch and still-face-with-touch pe-riods.

Discussion

As predicted, infants of depressed moth-ers responded more positively to the rein-statement of touch following a still-face-no-touch episode than did infants of nonde-pressed mothers. Infants of depressed moth-ers showed more positive affect (moresmiles and vocalizations) and gazed more attheir mothers' hands during the still-face-with-touch period than the infants of nonde-pressed mothers, who grimaced, cried, andgazed away from their mothers' face moreoften during this period.

We should note that, by specifically in-troducing a still-face-with-touch period im-mediately after a still-face-no-touch period,we were able to measure the soothing effectsof touch in the still-face situation right whenthe infants began showing the distressing ef-fects produced by their mothers' still-facewithout touch. All infants became somewhatsimilarly distressed during the still-face-no-touch procedure; in particular, their gazeaway from mothers' face (gaze aversion) wassignificantly higher during this period com-pared to the other three periods. When touchwas introduced in the still-face situation,

however, the effects were more soothing forthe infants of depressed mothers and gazeaversion significantly decreased, but onlydecreased for the infants of depressed moth-ers. In this way, the distress caused by ma-ternal lack of facial expressions and voicewas reduced by instructing mothers to ac-tively touch their infants. These findingssuggest that the effects caused by the stillface (lack of emotional expressions) can bepartially eliminated (or reduced) by mothersactively touching their infants while still fa-cially and verbally unresponsive.

In general, both groups of infantsseemed to like touch, and they showed it bysmiling and vocalizing more when theywere touched. This study extends the previ-ous findings (Pehiez-Nogueras et aI., 1996;Stack & Muir, 1990, 1992) by examining dif-ferences between depressed and nonde-pressed groups. In addition to finding thatinfants of depressed mothers smiled and vo-calized more, oriented more to their de-pressed mothers, and cried and grimacedless than infants of nondepressed mothersduring the still-face-with-touch period, wefound that during the final return to normalplay period infants of nondepressed mothersdid not appear to "recover" from the dis-tressing still-face periods, and they began tocry, grimace more, gaze away more, and tosmile and vocalize less compared to infantsof nondepressed mothers. Because we mini-mized the potential sociodemographic con-founds by having a homogeneous sample ofdepressed and nondepressed adolescentmothers of low SES, our results can be con-sidered representative for this particularlower-income adolescent population. Giventhis homogeneity of our sample, the resultsmay be limited in generalizability.

The effects observed in infant behaviorwere not accounted for by immediate groupdifferences in maternal behavior. That is,the depressed and nondepressed mothers'behaviors were not significantly different inthe conditions in which infant behavior dif-ferences were observed. This uniform pat-tern of maternal behavior was expectedgiven the highly standardized procedures ofthis study with both groups of mothers. Bothdepressed and nondepressed mothers werespecifically instructed and given a demon-stration showing them how to behave in thestill-face-no-touch period and how to touchtheir infants during the still-face-with-touchperiod. In the absence of immediate groupdifferences in maternal behavior, the differ-ences in the pattern of infant behaviorsacross conditions can be related to the in-

TABLE

2

MEA

NPER

CEN

TAG

EOF

MO

THER

S'(N=

16D

epressedand

N=

16N

ondepressed)BEH

AV

IORSA

CRO

SSFOU

RSU

CCESSIV

EPERIO

DS

NO

RM

ALPLA

YSTILL

FACE-N

o-ToUCH

STILL-FACE-W

ITH-TOU

CH

NO

RM

ALPLA

Y

Non-

Non-

Non-

Non-

Depressed

depressedD

epresseddepressed

Depressed

depressedD

epresseddepressed

Mother's

totaltouch:

63.865.3

.47.37

97.993.7

88.975.6

(31.3)(30.1)

(1.5)(1.0)

(3.4)(15.2)

(12.3)(27.1)

Resting

onbaby

............10.8

10.8.00

.1212.4

11.224.4

18.3(10.3)

(9.3)(.00)

(.50)(ILl)

(13.3)(18.7)

(19.4)M

ildtouch

.................24.4

16.4.47

.2547.8

54.321.3

11.6(24.7)

(17.4)(1.5)

(1.0)(30.4)

(36.0)(19.1)

(16.0)Intense

touch..............

7.1ILl

.00.00

13.29.0

12.17.3

(11.0)(13.8)

(.00)(.00)

(19.9)(12.7)

(11.8)(7.6)

Mild

movem

ent............

21.626.1

.00.00

24.519.3

27.834.6

(31.0)(20.2)

(.00)(.00)

(26.8)(24.1)

(24.2)(23.2)

Intensem

ovement

...........0

1.0.0

.0.0

.03.4

4.0(.0)

(2.2)(.0)

(.0)(.0)

(.0)(6.5)

(5.9)M

other'svocal

sounds.........

86.589.7

2.7.7

3.61.2

80.689.5

(13.4)(12.8)

(3.9)(1.9)

(9.5)(3.4)

(22.7)(16.5)

Mother's

face:Sm

iles.....................

70.466.2

4.32.8

3.1.9

51.846.7

(21.7)(32.0)

(7.3)(6.7)

(5.9)(2.2)

(28.8)(28.7)

Negative/angry

.............1.3

1.23.5

5.33.9

3.51.6

4.8(3.4)

(3.8)(14.0)

(20.7)(13.0)

(9.9)(3.5)

(9.6)N

eutral....................

28.432.8

92.391.7

93.195.6

46.748.7

(2Ll)(32.7)

(14.6)(21.5)

(14.2)(9.8)

(27.6)(27.9)

NO

TE.-MA

NO

VA

srevealedno

groupand

nogroup

Xperiod

interactioneffects.Standard

deviationsare

inparentheses.

fant's prior history of interactions with a de-pressed mother. The differences observed inthe infants' behavior may be attributed tomaternal depression and its concomitant his-tory of interactions.

Alternatives to learning-history explana-tions, however, should be considered. It hasbeen argued that infants of depressed moth-ers are at unusually high risk for developingdepression due to genetic or prenatal trans-mission (Zuckerman, Als, Bauchner, Parker,& Cabral, 1990). To predict infant "de-pressive" behavioral outcomes from any sin-gle factor, however, is almost impossible,whether this factor is genetic or postnatal be-havior experience. To elucidate the etiologyof infant depression was not the objective,rather, our goal was to determine if a shortintervention with touch by depressed moth-ers would increase infants' positive affectand attention during still-face interactions.As predicted, touch was more effective inenhancing the positive behavior in infantsof depressed mothers. Depressed mothersseemed to have facilitated more positive af-fect and attention in their infants by touch-ing them during the interactions, and the op-timallnonintrusive type of touch used in thisstudy appeared to provide comfort duringthe stressful still-face interactions.

The findings can be explained in a num-ber of ways. One possible explanation is thatthe infants of nondepressed mothers did notshow significantly less grimacing and cryingwhen touch was introduced during the stillface because these infants were less familiarwith maternal unavailability (flat face and af-fect) and were thus much more difficult tosoothe when touch was introduced. More-over, infant grimacing and crying continuedto be emitted by the infants of nondepressedmothers even during the resumed normalperiod. The increase in infant grimacing andcrying during the mothers' subsequent re-turn to normal play following a period ofma-ternal unavailability was also observed byToda and Fogel (1993).

Typically, mothers "fake good" and "tryharder" to show positive behaviors duringthe initial moments of videotaping in experi-ments. Thus, the absence of group differ-ences in maternal behavior in the initialnormal play period should not be overinter-preted. For the purpose of this study, thefirst normal play period may neither be rep-resentative of a "true" baseline nor as rele-vant as the subsequent differences observedlater on the final play period after still-faceperiods. It is possible that touch was moresoothing for infants of depressed mothers,

who might normally be deprived of contin-gent maternal touch and contact at home.Touch may have quickly become nonsooth-ing, and perhaps aversive, for the infants ofnondepressed mothers, who might normallynot be deprived and were more upset andstressed by the preceding still-face-no-touchperiod.

During the resumed normal play inter-action, then, the infants of depressed moth-ers were not as distressed as those of nonde-pressed mothers. Touch was initiallysoothing to the infants of depressed mothers,and in the aftermath of the still-face periods,the infants of nondepressed mothers weremore upset. This phenomenon suggests thatas a result of their history of experiences,the infants of depressed mothers were lessdistressed by the still-face perturbations andthe absence of maternal touch. Conceivably,infants of depressed mothers could havebeen less distressed in our study becausethey received more optimal touch than theywere used to.

The results of the present study can berelated to findings from a recent learningexperiment using a synchronized reinforce-ment procedure (Pelaez-Nogueras et aI.,1996). Pelaez-Nogueras and colleaguesfound that contingent tactile stimulation bya caregiver during face-to-face interactionsincreases affect and attention in 3-month-oldinfants. In that study, when touch was usedas part of the caregiver's social stimulationand provided contingently, it effectively re-inforced and maintained higher rates of in-fant eye contact, smiles, and vocalizations.Interestingly, the infants in the presentstudy also increased eye contact with theirmothers during the still-face-with-touch pe-riod; this may have occurred as a result ofintermittent contingent touch stimulationon infant making eye contact with theirmothers.

Both learning and emotional regulationprpcesses prepare the infant to develop adap-tive and organized behavior strategies (Pel-aez-Nogueras, 1992; Thompson, 1994). Thedifferences observed in the infants' behaviormay lie in the different histories of interac-tions between mother and child and historiesof infant behavior regulation. The data for thefirst normal play period show that infants ofdepressed and nondepressed mothers dif-fered in facial grimacing and looking athands, suggesting differences in their learn-ing histories and conceivably in their abilityto regulate their behaviors. However, eventhough there were differences between in-fants of depressed and nondepressed moth-

ers, we should be cautious when attributingthese differences to the infants' prior interac-tive histories with their mothers in light ofthefact that there were no immediate group dif-ferences in the mothers' behaviors in the firstnormal play period.

In sum, the effects of maternal touchduring still face were more powerful for in-fants of depressed mothers than for infantsof nondepressed mothers, even when theamount and type of touch provided by thedepressed and nondepressed mothers werethe same. Touch appears to have strong posi-tive influences on infant behavior, it can in-crease positive affect, increase infants' nega-tive affect, and direct infants' attention, inparticular, the attention of infants of de-pressed mothers during face-to-face interac-tions. The type of stimulation that involvestouch during face-to-face interactions needsto be investigated further. Although short-term positive effects were achieved in thepresent study, long-term assessments andimplementations of this type of interventionare needed to determine the more prolongedpositive effects of touch on infant behavior.Future research should focus on touch inter-vention strategies with infants and their de-pressed mothers.

Beardslee, W. R., Bemporad, J., Keller, M. B., &Klerman, G. L. (1983). Children of parentswith major depressive disorder: A review.American Journal of Psychiatry, 54,1254-1268.

Beck, A. T., Ward, C. H., Mendelson, M., Mach,J. E., & Erbaugh, J. (1961). An inventory formeasuring depression. Archives of GeneralPsychiatry, 4, 561-571.

Campbell, S. B., Cohn, J. F., & Meyers, T. (1995).Depression in first-time mothers: Mother-infant interaction and depression chronicity.Developmental Psychology, 31, 349-357.

Cohn, J. F., & Campbell, S. B. (1992, May). Thecourse and correlates of depression in post-partum women and their influence on infantsocio-emotional development. Symposiumdelivered at the International Conference onInfant Studies, Miami.

Cohn, J. F., Campbell, S. B., Matias, R., & Hop-kins, J. (1990). Face-to-face interactions ofpostpartum and nondepressed mother-infantpairs at two months. Developmental Psychol-ogy, 26, 185-193.

Cohn, J. F., Matias, R., Tronick, E. Z., Connell,D., & Lyons-Ruth, K. (1986). Face-to-face in-teractions of depressed mothers and their in-fants. In E. Z. Tronick & T. Field (Eds.), Ma-ternal depression and infant disturbance

(New Directions for Child Development, No.34). San Francisco: Jossey-Bass.

Cohn, J. F., & Tronick, E. Z. (1983). Three-month-old infants' reaction to simulated maternal de-pression. Child Development, 54, 185-193.

Cummings, E. M., & Cicchetti, D. (1990). Towarda transactional model of relations betweenattachment and depression. In M. T.Greenberg, D. Cichetti, & E. M. Cummings(Eds.), Attachment in the preschool years:Theory, research, and intervention (pp. 339-372). Chicago: University of Chicago Press.

Field, T. (1977). Effects of early separation, inter-active deficits, and experimental manipula-tions on infant-mother face-to-face interac-tion. Child Development, 48, 763-771.

Field, T. (1984). Early interactions between in-fants and their postpartum depressed moth-ers. Infant Behavior and Development, 7,527-532.

Field, T. (1986). Models for reactive and chronicdepression. In E. Tronick & T. Field (Eds.),Maternal depression and infant disturbance(pp. 47-60). San Francisco: Jossey-Bass.

Field, T. (1992). Infants of depressed mothers. De-velopment and Psychopathology, 4, 49-66.

Field, T., Healy, B., Goldstein, S., & Guthertz, M.(1990). Behavior state matching in mother-infant interactions of nondepressed versusdepressed mother-infant dyads. Develop-mental Psychology, 26, 7-14.

Field, T., Schanberg, S. M., Scafidi, F., Bauer,C. R., Vega-Lahr, N., Garcia, R., Nystrom, J.,& Kuhn, C. M. (1986). Tactile/kinestheticstimulation effects on preterm neonates. Pedi-atrics, 77, 654-658.

Field, T., Vega-Lahr, N., Scafidi, F., & Goldstein,S. (1986). Effects of maternal unavailabilityon mother-infant interactions. Infant Behav-ior and Development, 9, 473-478.

Gaensbauer, T. J., Harmon, R. J., Cytryn, L., &McKnew, D. H. (1984). Social and affectivedevelopment in infant with manic-depressiveparent. American Journal of Psychiatry, 141,223-229.

Gelfand, D. M., & Teti, D. M. (1990). The effectsof maternal depression on children. ClinicalPsychology Review, 10, 329-353.

Gewirtz, J. L., & Pehiez-Nogueras, M. (1992). In-fant social referencing as a learned process.In S. Feinman (Ed), Social referencing andthe social construction of reality in infancy(pp. 151-173). New York: Plenum.

Goodman, S. (1992). Understanding the effects ofdepressed mothers on their children. In E. F.Walker, B. Cornblatt, & R. Dworkin (Eds.),Progress in experimental personality andpsychopathology research (Vol. 15, pp. 47-109). New York: Springer.

Gusella, J. L., Muir, D. W., & Tronick, E. Z.(1988). The effect of manipulating maternal

behavior during an interaction on 3- and 6-month-olds' affect and attention. Child Devel-opment, 59, 1111-1124.

Guthertz, M., & Field, T. (1989). Lap computer oron-line coding and data analysis for laboratoryand field observations. Infant Behavior andDevelopment, 12,305-319.

Hammen, C. (1992). Cognitive, life stress, and in-terpersonal approaches to a developmentalpsychopathology model of depression. Devel-opment and Psychopathology, 4, 189-206.

Lamb, M. E., Morrison, D. C., & Malkin, C. M.(1987). The development of infant social ex-pectations in face-to-face interaction: A longi-tudinal study. Merrill-Palmer Quarterly, 33,241-254.

Lyons-Ruth, K., Zoll, D., Connell, D., & Grune-baum, H. U. (1986). The depressed motherand her one-year-old infant: Environment, in-teraction, attachment, and infant develop-ment. In E. Z. Tronick & T. Field (Eds.), Ma-ternal depression and infant disturbance(New Directions for Child Development, No.34). San Francisco: Jossey-Bass.

Malphurs, J., Raag, T., Field, T., Pickens, J., &Pelaez-Nogueras, M. (1996). Touch by intru-sive and withdrawn mothers with depressivesymptoms. Early Development and Parent-ing, 5, 122.1-5.

Mayes, L. C., & Carter, A. S. (1990). Emergingsocial regulatory capacities as seen in thestill-face situation. Child Development, 61,754-763.

Ottenbacher, K. J., Muller, L., Brandt, D., Heint-zelman, A., Hojem, P., & Sharpe, P. (1987).The effectiveness of tactile stimulation as aform of early intervention: A quantitativeevaluation. Developmental and BehavioralPediatrics, 8, 68-76.

Pelaez-Nogueras, M. (1992). Infant learning toreference maternal emotional cues. Unpub-lished doctoral dissertation, Florida Interna-tional University.

Pelaez-Nogueras, M., Field, T., Cigales, M., Gon-zalez, A., & Clasky, S. (1994). Infants of de-pressed mothers show less "depressed" be-havior with their nursery teachers. InfantMental Health journal, 15, 358-367.

Pelaez-Nogueras, M., Gewirtz, J. L., Field, T., Ci-gales, M., Malphurs, J., Clasky, S.,& Sanchez,A. (1996). Infants' preference for touch stimu-lation in face-to-face interactions. journal ofApplied Developmental Psychology, 17, 199-213.

Sameroff, A. V., & Seifer, R (1983). Familial riskand child competence. Child Development,54, 1254-1268.

Scafidi, F. A., Field, T., Schanberg, S. M., Bauer,C. R, Tucci, K., Roberts, J., Morrow, C., &Kuhn, C. M. (1990). Massage stimulatesgrowth in preterm infants: A replication. In-

fant Behavior and Development, 13,167-188.

Stack, D. M., & Muir, D. W. (1990). Tactile stimu-lation as a component of social interchange:New interpretations for the still-face effect.British journal of Developmental Psychol-ogy, 8, 131-145.

Stack, D. M., & Muir, D. W. (1992). Adult tactilestimulation during face-to-face interactionmodulates five-month-olds' affect and atten-tion. Child Development, 63, 1509-1525.

Teti, D. M., Gelfand, D. M., Messinger, D. S., &Isabella, R (1995). Maternal depression andthe quality of early attachment: An examina-tion of infants, preschoolers, and their moth-ers. Developmental Psychology, 31, 1-13.

Thompson, R (1994). Emotion regulation: Atheme in search of definition. In N. A. Fox(Ed.), The development of emotion regula-tion: Biological and behavioral consider-ations. Monographs of the Society for Re-search in Child Development, 59(2-3, SerialNo. 240).

Toda, S., & Fogel, A. (1993). Infant response tothe still-face situation at 3 and 6 months. De-velopmental Psychology, 29, 532-538.

Tronick, E. Z., Als, H., Adamson, L., Wise, S., &Brazelton, T. B. (1977). The infant's responseto entrapment between contradictory mes-sages in face-to-face interaction. journal ofChild Psychiatry, 17, 1-13.

Tronick, E. Z., & Gianino, A. F. (1986). The trans-mission of maternal disturbance to the infant.In E. Z. Tronick & T. Field (Eds.), Maternaldepression and infant disturbance (New Di-rections for Child Development, No. 34). SanFrancisco: Jossey-Bass.

Weissman, M. M., Prusoff, B. A., Gammon, G. D.,Merikangas, K. R, Leckman, J. F., & Kidd,K. K. (1984). Psychopathy in children (ages6-18) of depressed and normal parents. jour-nal of American Academy 'of Child Psychia-try, 23, 78-84. .

Whiffen, V. E., & Gottlib, I. M.'(1989). Infants ofpostpartum depressed mothers: Tempera-ment and cognitive status. Journal of Abnor-mal Psychology, 98, 274-279.

Winer, B. J. (1971). Statistical principles in exper-imental design (2d ed.). New York:McGraw-Hill.

Zahn-Waxler, C., Cummings, E. M., McKnew,D. H., & Radke-Yarrow, M. (1984). Problembehaviors and peer interactions of young chil-dren with a manic-depressive parent. Ameri-can journal of Psychiatry, 141,236-240.

Zuckerman, B., Als, H., Bauchner, H., Parker, S.,& Cabral, H. (1990). Maternal depressivesymptoms during pregnancy, and newborn ir-ritability. Developmental and Behavioral Pe-diatrics, 11, 190-194.