Poster group
Details of individual items:
poster
Based on evidence that odors are a salient feature in young infants' environments, it was hypothesized that odors would serve as an effective contextual cue for memory retrieval in 2-month-old infants. As in prior studies with visual and auditory cues, it was expected that after a 1-day retention interval, infants would demonstrate retention regardless of the similarity between odors present during training and testing.Twenty-five middle class, apparently healthy, Caucasian infants who were an average of fifty-seven days old (sd 2.5) learned the mobile conjugate reinforcement task at home on 2 consecutive days and their memory was tested 1-day later. During each session, infants were first exposed to a non-moving mobile hanging above the crib for 3 min. Then, infants were attached via a ribbon tied around their ankle to the mobile such that their kicks moved the mobile for 9 min. Finally, infants were disengaged from the mobile and exposed to the non-moving mobile for 3 min. During Session 1, the first 3 min served as a baseline. During Session 3, the first 3 min served as a long-term retention test.Infants were randomly assigned as they became available to one of four groups: (1) infants who were trained and tested with the same odor infused in their nursery; (2) infants who were trained with one odor but tested with a different one; (3) infants who were trained with an odor but tested with no odor; and (4) infants who received no odor during either training or testing. The odors were either a floral or a woody scent.A 'baseline ratio' for each infant was computed by dividing response rate during the long-term retention test phase of Session 3 by the pretraining baseline response rate from Session 1. Baseline ratios above 1 indicate that the task has been remembered. Consistent with previous research, the baseline ratios obtained were positively skewed, necessitating a logarithmic transformation prior to analysis.The log-transformed baseline ratios were subjected to a series of directional t-tests to compare each group's mean log baseline ratio to a theoretical baseline of 0.00, which indicates a return to the pretraining operant level (log 1 0). These analyses indicate whether the infants in any given group have remembered the task. A log-baseline ratio significantly greater than 0.00 was obtained only for the same odor group, t(5) 3.3, p .02.In the present study with 2-month- infants, unimpaired retention was observed only in the same-odor group. Surprisingly, retention was not seen in the no-odor group. Thus, having the same odor during training and testing facilitated retention.These findings are similar to the findings of Rubin, Fagen, and Carroll's (1999) olfactory experiments with 3-month-old infants in which unimpaired retention was observed only in the same odor group and only when the retention interval was 1-day. They are also consistent with studies of odor-context effects in adults where it has been found that a training/testing match in odor facilitates retention after delays of 48 hr or less.
poster
Olfactory learning by neonates has been demonstrated in experiments using avariety of training/testing procedures, includinghabituation/dishabituation, mere exposure, classical conditioning and otherforms of associative learning. In previous research on the effects of mereexposure, infants displayed preferential responses to an odor to which theyhad been exposed for several (19-24) hours during the first 2 days afterbirth. The present experiment is an initial attempt to determine whethernewborns would become familiar with (and subsequently respond preferentiallyto) an odor that they experience for only a brief period immediately afterbirth. Moreover, all subjects had been delivered via Caesarean section (CS).These babies are often separated from their mothers after birth and cantherefore experience different non-natural odors. It is also known that CSdelivery influences newborns' perceptual capabilities and behavior: e.g.altered endocrine activity due to a short or absent birth stress relative toinfants born vaginally. Eighteen healthy, term CS babies (11 boys, 7 girls) participated inthe experiment. Ten minutes after birth, a gauze pad treated with aromaticoil of cherry or passion fruit was placed 5-cm from the baby's nose andremained there throughout the 30-min. exposure session. All babies weretested 2-5 days later for their responses to cherry vs. passion fruit odorspresented on 2 gauze pads simultaneously suspended along each side of theirface. During 2 successive 60-sec. tests, the babies spent significantly moretime with their face turned towards the exposure odor than to the unfamiliarcomparison odor. In a second series of tests, the same newborns alsooriented for a reliably longer time to the exposure odor than to a cleancontrol pad. Our data further support the conclusion that newborn infants arehighly sensitive to environmental odors. Thirty minutes of exposure to anodorant within the first hour following birth was sufficient for newborns tobecome familiar with that scent and remember it several days later. Thephysiological and sensorial concomitants of CS delivery do not appear tointerfere with early olfactory perception and learning.
poster
When adults are asked to report their earliest personal experiences,most can recall little about events that occurred prior to the age of3 or 4 years. The absence of memories for our infancy and earlychildhood is commonly referred to as infantile or childhood amnesia.The pervasive finding that our early life is veiled from recollectionraises basic questions about the nature of memory processing duringinfancy and early childhood. Historically, the period between 18- and24-months of age has marked an important transition in cognitivedevelopment. According to Piaget (1962), this developmental epochrepresents the final stage of the sensorimotor period and marks theemergence of representational thought. Although some contemporarydevelopmental psychologists have rejected many aspects of traditionalPiagetian theory, many still support the notion that the periodbetween 18 and 24 months is characterized by a fundamental shift inmental representation that may have important implications for memorydevelopment. Furthermore, the 'vocabulary spurt' that commonly occursduring this period may alter the way in which infants can encode,retrieve, and express their memories. In the present experiment weexamined age-related changes in memory performance between 18 and 24months of age. To do this, 18- and 24-month-old infants were visitedin their own home on two occasions. During the initial visit,children's receptive and productive language skills were measured. The experimenter then showed the child how to operate a special a boxthat changed old toys into new toys. First, the experimenter showedthe child how to place a toy in a hole at the top of the box. Next,the child was shown how to press a button on the front panel of thebox that produced music. Finally, the child was shown how to open adoor at the bottom of the box to retrieve a new toy. This sequencewas repeated a total of nine times. Each infant was tested 24-hourslater. During the test, both verbal recall and behavioral re-enactmentwere assessed. Consistent with prior research using imitation and re-enactment procedures, infants of both ages exhibited excellent retention on measures of behavioral re-enactment. In contrast, infants of both ages exhibited virtually no retention on measures of verbal recall, despite the fact that the words necessary to answer the questions were part of their productive vocabulary (e.g., ball, teddy). Given this, we hypothesize that children's performance on the verbal recall task reflects the nonverbal nature of their memory representation rather than an inability to express information that has been encoded in a verbal way. Prior to language acquisition, for example, infants must encode information in a nonverbal way. Even as their language skills improve, however, young children rely more heavily on nonverbal representation than on their somewhat limited language skill. As a test of this hypothesis, we are currently assessing infants' behavioral and verbal recall of this task after longer delays (e.g., 3 and 6 months) to determine if their content of their verbal reports increase as their language skills continue to improve.
poster
One hallmark of memory development during the infancy period is anage-related increase in the range of effective retrieval cues for aparticular memory. Studies conducted with 2- to 6-month-old infants,for example, have shown that memory retrieval is precluded by a changein the environmental context at the time of the test. As a functionof age and experience, however, infants begin to retrieve and expresstheir memories in a wider range of situations. By 12 months of age,for example, memory retrieval is unaffected by a change in theenvironmental context at the time of the test. Rovee-Collier (1997) has recently argued that the onset of independentlocomotion may mark an important transition in memory development. According to her theory, the importance of contextual similarity formemory retrieval decreases as infants learn to crawl. Alternatively,Bertenthal and Campos (1990) have argued that although the onset ofindependent locomotion plays an important role in object permanenceand social referencing, it does influence recognition memory,imitation, or response to novelty. The purpose of the presentexperiment was to examine the relation between independent locomotionand the flexibility of memory retrieval using a deferred imitationprocedure.A total of 90 9-month-old infants participated in the presentexperiment. At the time of their participation, half of the infantswere crawling and half were not. An equal number of crawlers andnon-crawlers were assigned to the demonstration or control conditions. Infants in the demonstration condition watched as an experimenterperformed a single target action with a novel object. Independentgroups of infants were tested 24 hours after the demonstration withthe same object in the same context (no change), a different object inthe same context (cue change), the same object in a different context(context change), or a different object in a different context (cueand context change). To assess the spontaneous production of targetactions in the absence of an adult demonstration, infants in thecontrol condition were not shown the target action or the object priorto the test. When tested after a 24-hour delay, a change in the cue or a change inthe context had no effect on memory performance. The imitation scoresof infants in the cue change and context change test groups were notdifferent from those of infants in the no change group and weresignificantly higher than those of infants in the control condition. When both the cue and the context were changed at the time of thetest, however, retention was impaired. Importantly, there was nodifference in memory performance as a function of the infants'crawling status. Taken together, the present findings suggest thatthe onset of independent locomotion may not play a major role in theflexibility of memory retrieval, at least when infants are tested in adeferred imitation procedure after only a 24-hour delay.
poster
ERP waveform components (Pb, Nc and Nc2) from infants experiencing a visual recognition memory paradigm are superimposed upon the early negative slow wave (NSW) which may be expressed pre-stimulus and typically no later than 10-50 ms after stimulus onset. The early onset of NSW may be related to the ability of infants to express expectancy (Wentworth et al., 1998) or anticipatory attention for specific stimuli (Karrer and Ackles, 1990). Significant interaction effects of experience on NSW between infants with and without DS (Hill Karrer et al., 1998) supports this hypothesis. The current study was designed to test NSW expression as an ability of infants to utilize sequential information processing to form expectations concerning upcoming stimuli in order to detect an infrequent stimulus omission. Two infant-controlled paradigms were utilized. In both, a total of 80 stimuli were presented. Two smiling Caucasian female faces were presented as stimuli for 80% frequent/ 10% novel probabilities. Stimuli were counterbalanced across infants. The remaining 10% of stimuli were omitted. The first stimulus omission paradigm involved no temporal clue that a stimulus was omitted (SO-B). That is, a completely black screen, identical to the 'blank' screen during an ISI, was presented during SO-B. The second paradigm prompted a neutral gray screen during the stimulus omission (SO-G), giving temporal representation of 'omission'. Infants with DS (n 3D 13), FXS (n3D5) and typically developing infants (n 3D 26) participated in the study. The complexity of the paradigm and the relatively short exposure of novel stimuli prompted significant differences in fixation behaviors and ERPs between study groups. Typically developing infants demonstrated significantly longer fixations in response to the presentation of novel stimuli. Inversely, infants with DS demonstrated less fixation duration in response to novel stimuli. As a result, there was a significant interaction of fixation response to novel stimuli between the two study groups. All infants within the three study groups prolonged their fixations after SO-G. However, only the typically developing infants detected SO-B and prolonged their fixations after SO-B. Behavioral data are supportive of ERP data that were simultaneously recorded in the same paradigms. During frequent stimuli, infants with DS expressed significantly larger and infants with FXS significantly smaller ERP components than typically developing infants. Among typically developing infants, probability effects were robustly recorded. However, there were few ERP probability effects expressed by infants with DS and FXS. Moreover, there was no indication of change from baseline EEG during SO-B among infants with DS and FXS. Thus, both behavioral and neurophysiological data demonstrate that DS and FXS study groups did not detect novel facial stimuli, or SO-B. Among typically developing infants, a slow negative wave was expressed during SO-B, but only at Fz. NSW during SO-B was especially prominent among infants that did not avert their gaze during the long ISI within stimulus omission. Clearly, typically developing six-month-old infants are capable of complex cognitive processes, and simultaneous measures of ERPs and behavior illuminate these remarkable abilities. These studies therefore imply that NSW provides a measure of expectancy and anticipatory executive functions during infancy, and that NSW is capable of illustrating delayed development within prefrontal functions at an early age.
poster
Previous studies found that auditory event-related potentials (ERPs) recorded to familiar and novel tones from 5-month-old infants had larger amplitudes and less variability to the familiar tones (familiarization occurred 24 hours previously). These results were quite robust and occurred at both frontal and central electrode sites for a positive ERP component with a latency of approximately 250 ms (P2) and a later negative component at 400 ms (N2). Two subsequent studies with 3-month-olds found similar familiar/novelty differences but which were considerably weaker. We posited that the differences between 3- and 5-month-olds were due not to qualitative developmental differences between the 2 age groups, but to the amount of general experience with tones that each age group had had coming into the experiment. Therefore, we proposed that 3-month-olds should show robust differences between familiar and novel stimuli similar to 5-month-olds when linguistic stimuli are used, i.e., when the class of stimuli is more familiar to them. On Day 1 of testing, 100 presentations of a single phoneme ('ba' or 'da') were given over headphones to 32 three-month-old infants. The following day, subjects received 50 presentations of the familiar phoneme and 50 presentations of a novel phoneme. ERPs were recorded from prefrontal (Fpz), frontal (Fz), and central (Cz) locations. Traditional average ERP amplitudes were measured for P2 and N2 at each electrode. In addition, a template-matching procedure using the average ERP as the template was utilized to estimate the amplitude and latency of P2 and N2 in each single trial. The mean of these amplitudes was thus used as a second measure of amplitude, called single-trial amplitude. The standard deviation of the single-trial latencies was used as a measure of the temporal variability of the ERPs. A priori hypotheses predicted that average and single-trial amplitude would be larger for Day 2-Familiar (D2F) compared to Day 2-Novel (D2N), and that D2F latency variability would be less than that of D2N. Planned comparisons utilizing Bonferroni-protected t-tests revealed statistically significant differences only for the N2 peak at the Cz electrode. Average ERP amplitude was significantly larger for D2F than D2N (-6.7 (V and -4.8 (V, respectively), t(31) 2.3, p < .05. Single-trial amplitude also followed this trend (D2F -14.8 (V; D2N -10.9 (V), t(31) 2.2, p .05. Finally, D2F latency variability was less than that of D2N (80.2 ms and 85.6 ms, respectively), t(31) 2.5, p < .05.These results confirm previous findings that ERPs from young infants show increased amplitude and decreased latency variability to familiar stimuli, which can be interpreted to mean that stimulus experience serves to stabilize the neural ensemble that encodes and stores that stimulus. However, these results disconfirm our hypothesis that linguistic stimuli would result in more robust differences between familiar and novel that was previously found with 3-month-olds using pure tones. Aided by Social and Behavioral Sciences Grant No. 12-FY97-0115 from the March of Dimes Birth Defects Foundation.
poster
Gender and the societal roles attached to females and males play a major role in our lives. Adults and children use their gender knowledge and attitudes to guide behavior and attention, and to interpret and structure information (i.e., gender-based schematic processing; GBSP). An example of GBSP is memory distortion. When preschool-aged children remember gender-stereotype inconsistent information, they tend to distort, or mis-remember, that information, especially when the memory load is high2E For example, when 5- to 6-year-old children recall schema-inconsistent pictures (such as a girl sawing wood) after a one-week delay, they distort the sex of the actor. That is, they remembered the picture as a boy sawing wood (Martin & Halverson, 1983). However, when the memory load is smaller, gender-inconsistent information is not distorted. Seven- to eight-year-old children were able to remember stereotype-inconsistent pictures correctly when the task included no delay (Liben & Signorella, 1980).Until recently, studies of children's GBSP were restricted to children ages three and up. The measures in these studies were primarily verbal, and beyond the skills of children under three years. Using age-appropriate methodologies with pre- and early-verbal infants makes it possible to extend the age range in which GBSP is studied. Recent studies have found GBSP in children under the age of three years (e.g., Bauer, 1993; Sen & Bauer, 1999). In addition, this age range is a critical one from a theoretical perspective. Gender schema theories propose that GBSP begins when gender becomes a salient distinction for infants (Martin & Halverson, 1981; Poulin-Dubois et al., 1994). Although the type of knowledge that makes gender salient is still not determined, the posited abilities all develop before the age of three. Thus, it is important to study GBSP in these younger children. The youngest age at which memory distortion has been demonstrated is 3 years (Cordua et al., 1979). However, it seems likely that younger children might also distort stereotype-inconsistent material. In the current study, 24-month-old infants' memory for counter-stereotypical activities is being tested using the inductive generalization paradigm. This paradigm has been used to test children's categorical knowledge during the preverbal period (McDonough & Mandler, 1998), and has been adapted to test infants' gender-stereotype knowledge (Poulin-Dubois, Serbin, Eichstedt, Sen, & Biessel, 1999). With this paradigm, infants are presented with a toy (e.g., a vacuum) in the baseline, familiarization stage. Next, an activity (e.g., vacuuming) is modeled by the experimenter (modeling stage) using a gender-neutral doll (toy monkey). The infants are then presented with a female and a male doll and encouraged to imitate the activity (generalization stage). By 31 months, girls and boys show some gender-stereotype knowledge (e.g., imitating the feminine activities with the female doll).Gender-stereotyped activities (e.g., putting on makeup) are modeled using the gender-'inappropriate' doll. Children are then presented with different female and male dolls. Half of the children have a delay between the two stages, and half imitate the activities immediately. Preliminary pilot data suggest that 24-month-old infants show distorted memory of the cross-gendered activities. A second group of infants will participate in the control condition, in which their knowledge of the stereotyped nature of these particular activities will be tested using the procedure used by Poulin-Dubois and colleagues.This study is currently in progress, but will be completed before the conference. A younger age group may also be tested.
poster
In the last decade, three paradigms have emerged as popular methods ofmeasuring infant memory processing. In the deferred imitation paradigm,infants observe an experimenter perform a series of novel actions with anobject and their ability to reproduce those actions is assessed following adelay. In the mobile conjugate reinforcement paradigm, infants learn tokick their feet to produce movement in an overhead crib mobile and theirkick rate is re-assessed following a delay. In the visual recognitionmemory (VRM) paradigm, infants are familiarized with a stimulus and theirlooking time toward that stimulus relative to a novel one is comparedfollowing a delay.Recently there has been considerable debate about the comparability ofthese three paradigms. It has been argued, for example, that the deferredimitation and VRM paradigms provide a measure of declarative memoryperformance while the mobile conjugate reinforcement paradigm provides ameasure of nondeclarative memory performance. Recent reviews of theliterature, however, suggest that the same experimental variables thatinfluence infants' performance in one paradigm (e.g., effects of contextchange, delay, and age) also influence performance in the others. Thepurpose of the present study was to systematically compare infants'performance across all three paradigms.In Experiment 1, we compared 6-month-old infants' memory performance in thedeferred imitation and VRM paradigms. To do this, an experimenterdemonstrated three target actions with a hand puppet. These actions weredemonstrated a total of three times. Immediately after the demonstration,infants were simultaneously presented with 2 puppets; the demonstrationpuppet and a novel puppet. The amount of time that infants spent lookingat each puppet was recorded. Infants were then tested for imitation of thetarget actions with the demonstration puppet.In Experiment 2, we compared 6-month-old infants' memory performance in themobile conjugate reinforcement and VRM paradigms. To do this, infants wereinitially trained to kick their feet to produce movement in an overheadmobile. During training, each infant was placed in a sling seat inside aportable crib and a ribbon was tied from his or her ankle to a mobilesuspension bar. Training began with a 2-minute nonreinforcement phase(baseline) followed by a 6-minute reinforcement phase (acquisition).Immediately following the acquisition phase, infants were placed on theircaregiver's lap and were simultaneously presented with 2 mobiles; thetraining mobile and a novel mobile. The amount of time that infants spentlooking at each mobile was recorded. Infants were then placed back in thesling seat inside the crib and their kick rate was assessed in the presenceof the training mobile.Infants exhibited perfect retention when tested in the deferred imitationand mobile conjugate reinforcement paradigms, however, the same infantsexhibited no retention whatsoever when tested in the VRM paradigm.Contrary to current claims, measures of infant memory obtained usingimitation and conditioning procedures may actually be more similar thanthose obtained using imitation and visual recognition procedures. Theimplications of the present findings for current theories of memorydevelopment will be discussed.
poster
Recent studies have provided evidence of memory by 1- to 2-year-oldsafter delay intervals of as many as 4 months (e.g., Hudson & Sheffield,1998; Meltzoff, 1995). With few exceptions (e.g., Bauer, Hertsgaard, &Dow, 1994), retention over longer intervals has not been examined. Thus,whether the length of time over which young children are able toremember even begins to approximate the long intervals over whichpreschool age and older children remember (e.g., Hamond & Fivush, 1991)is an open question. Moreover, how the length of time over whichactivities are remembered is influenced by the type of activity has notbeen examined. To begin to establish the parameters of young children’slong-term memories, and the factors that influence retention, we testedmemory for different types of activities, over a range of delays. Participants were 150 normally developing children taking part in alongitudinal study. One-third of the children were enrolled at 13months, 1/3 were enrolled at 16 months, and 1/3 were enrolled at 20months. At each of 3 exposure sessions, spaced 1 week apart, childrenengaged in two activities: (1) silly bunny – putting a number ofstacking rings on the ears of a stuffed rabbit, and (2) Mr. Giraffe –locating and “greeting” a large giraffe decal hidden from view by acurtain. Each activity was initiated and led by an experimenter: Theexperimenter provided the label for the activity and then invited thechild to participate with her to “make a silly bunny,” and to help her“guess” who was behind the curtain and then wave to and “say hello toMr. Giraffe.” After delays of 1, 3, 6, 9, or 12 months (between-subjects; n 10 perdelay), the children returned to the laboratory for a memory test.Unlike at the exposure sessions, the activities were not led by theexperimenter. Instead, the experimenter asked the child to engage in theactivity on her/his own. For example, the experimenter said “Last timeyou were here we made a silly bunny. Show me how you make a sillybunny.” At the first exposure session, none of the children spontaneouslyengaged in either of the target activities; at all exposure sessions,the children readily participated in the activities once they wereinitiated by the experimenter. At delayed testing, across tasks anddelays, 20-month-olds engaged in a larger proportion of the targetactivities (M .60) than did 16-month-olds (M .35) and 13-month-olds(M .25); performance of the 16- and 13-month-olds did not differ.Overall, children evidenced high levels of retention of the actions formaking a silly bunny (Table 1): After 1 month, 87% of the childrenproduced the actions either spontaneously or after the experimenter’sverbal reminder; after 12 months, whereas none of the childrenspontaneously produced the actions, 37% performed the target actionsafter the reminder. The children’s location memory was characterized bya different forgetting function. After 1 month, 60% of the childrenevidenced memory for the location of Mr. Giraffe. By 6 months, only 10%of the children provided any evidence of location memory. After 12months, none of the children provided evidence of location memory. Therewere no interactions with age. The present study represents a test of long-term memory by childrenaged 13 to 20 months. Children evidenced memory for specific actionsperformed after delay intervals of as many as 12 months. In contrast,location memory was at virtually undetectable levels by 6 months. Theresults highlight the importance of task variables in children’slong-term retention, and indicate that memory for actions performed ismore robust than is location memory. Table 1Percentage of Children Showing Evidence of Memory for Each TargetActivity at Each Delay Interval Target ActivityDelay Interval Silly bunny Mr. Giraffe 1 month 87% 60% 3 month 45% 39% 6 month 53% 10% 9 month 36% 3%12 month 37% 0%
poster
Recent research suggests that the development of long-term declarative memory begins to coalesce between 9-12 months of age. In studies of 9-month-olds, researchers have found individual differences in performance on deferred imitation, a nonverbal analog to verbal recall. About half of infants imitate events after a delay, whereas the rest do not (Carver & Bauer, 1999; Meltzoff, 1988). This pattern of results could be indicative of a transition permitting retention of events over long periods. In the present study, we provide converging evidence from neural and behavioral measures of declarative memory as evidence that the memory systems of infants undergo significant development near the end of the first year of life. Recognition memory was tested using Event Related Potentials (ERPs) and recall memory was tested using deferred imitation. In particular, we sought to understand why half of the infants from previous studies failed to recall the events after a delay: 1) Did they fail to encode them, or 2) Did infants successfully encode the events and then forget them over the delay? To determine whether differential recall is due to encoding and/or consolidation differences, ERPs were assessed immediately after exposure and after a 1 week delay. During 3 visits over the course of one week, infants were exposed to novel events in a typical deferred imitation paradigm. During each session, infants observed the events being modeled but were not permitted to manipulate the props or imitate the events. At the end of the 3rd exposure session, infants' recognition memory was tested using ERPs. ERPs were recorded in response to still images of a familiar event and an event they have never seen. After one week, ERP testing was repeated with a different familiar event and a different novel event. One month later, infants returned to the laboratory to see if they could reproduce the event sequences of events they had seen before and of events they had never seen (as a within subject control). We hypothesized that ERPs obtained immediately after infants see the events modeled provide a test of whether infants encoded them; ERPs recorded one week later test recognition memory during or after consolidation of the event into long-term memory.As in previous studies, two groups of infants could be identified: those who could and those who could not recall the event sequences. The ERP data collected immediately after the 3rd exposure revealed that for infants who later recalled a complete sequence ('recallers'), brain activity differentiated the new from the old events; for infants who did not later recall ('non-recallers'), no differentiation was apparent. At this point, it appears as though 'non-recallers' failed to encode the events despite repeated exposure. However, the ERP data collected 1 week later do not support this conclusion. Surprisingly, the brain activity of non-recallers looked more similar to that of the recallers; both groups displayed evidence of differentiation of novel from familiar events. Rather than a complete failure to encode the events, it appears as though the non-recallers simply required more time to solidify their memory trace of the events. Although recognition memory for the familiar events was present for both groups at the time of the 2nd ERP test, for the non-recallers, after the 1 month delay, the memory trace was not sufficient to permit recall. In contrast, infants who subsequently recalled created a well integrated memory trace early on that was evident at both the immediate and delay ERP tests and remained for the test of recall memory. Thus, neural correlates of recognition memory were related to later behavioral indices of recall memory. This research represents an important step in examination of the emergence and early development of long-term declarative memory.
poster
The purpose of the present research was to assess the ability of 17- and20-month-old children to integrate information presented over time inorder to form a coherent representation of an event. Temporalintegration is presumed to provide a test of working memory abilitybecause a child must be able to keep one piece of information active inmemory long enough to encode the next relevant piece of information inorder to put the separate bits of information together.Elicited imitation involves using props to model event sequences made upof a series of distinct steps. After modeling, the props are given tothe child who is encouraged to imitate the sequence. To assess temporalintegration, the steps of several different event sequences either wereintermingled with one another during presentation or were presented overan extended period of time by imposing a delay between the modeling ofeach step.In Experiment 1, three 3-step event sequences were modeled to 17- and20-month-olds in the following manner: A1(sequence A, step 1), B1, C1,A2, B2, C2, A3, B3, C3. In one condition each step was modeled only onetime while in the other condition each step was modeled three times insuccession. After modeling, for each sequence in turn, necessary propswere provided to the child, who was encouraged to perform the sequence.20-month-olds performed better (measured by the number of target actionsand pairs of correctly ordered actions) than 17-month-olds, andperformance was better with three repetitions than with one. The meansfor 17-month-olds with three repetitions were consistently lower thanthe means for 20-month-olds with one.Experiments 2 and 3 tested the parameters of temporal integrationability in 17- and 20-month-old children. The goal of Experiment 2 wasto simplify the temporal integration task in order to assess this skillin 17-month-olds. The new conditions were three 2-step events presentedA1, B1, C1, A2, B2, C2, and one 3-step sequence presented with delaysbetween steps (i.e., A1…A2…A3). All conditions used three repetitionsof each step. Experiment 3 addressed the robustness of temporalintegration skill in 20-month-olds by adding distraction and longerdelays between steps in conditions otherwise identical to theone-repetition condition in Experiment 1.The results of Experiment 2 revealed no significant differences foreffects of condition. 17-month-olds did not perform better in the2-step or delay conditions than in the replication of Experiment 1.However, the means in the replication condition were higher than thosein Experiment 1, implying that 17-month-olds displayed greater temporalintegration in Experiment 2.In Experiment 3, no significant effects of condition were observed.20-month-olds performed equally well in the distraction and delayconditions as in the replication condition.The present research demonstrated that 17-month-olds have some abilityto temporally integrate information modeled in the elicited imitationformat. In addition, 20-month-olds demonstrate this ability even underadverse conditions including distraction and increased time strain onworking memory.
poster
The present study used a span task adapted from one previously used with older children to examine the effects of span length, age, and serial position on visual recognition memory in 5-, 7-, and 12-month-old infants. Participants were part of a larger ongoing longitudinal study of the development of specific cognitive abilities and processing speed in full-terms and preterms from infancy through age 3 years. The sample at 5, 7, and 12 months consisted of 146, 140, and 126 full-terms, respectively, and 49, 56, and 56 preterms (some preterms were newly enrolled at 7 months). Infants were shown brightly colored, visually interesting objects, presented sequentially in spans of one, two, three, and four, with a fixed period of familiarization for each object (10s at 5 and 7 months, 3 s at12 months). When all objects in a span had been presented, they were shown again in the same sequence but now paired with new ones of equal attractiveness for a test period of 10 s per pair. The percentage of test time spent looking at the new object was computed for each pair. Novelty percentages were averaged across all problems to yield a summary score and across problems within each span to yield span scores. Because preterm status was unrelated to performance, data are presented combined across groups. Mean novelty scores declined linearly with increasing span length at 5 and 7 months, F(1,194) 125.6 and 118.2, respectively, both p < .001, dropping from around 63% with a span of one to chance levels (49%) at a span of four. At 12 months, performance was above chance on all spans, with novelty scores, ranging from 56% to 61%, and there was no linear decline. These findings were confirmed in the subset of infants having data at all 3 ages: An age by span MANOVA yielded a significant improvement in performance with age, Wilks' ( .89, F(2,151) 9.19, p < .001, and a significant age x span interaction, ( .63, F(6,147) 14.56, p < .001. The interaction indicated that the consistent performance across spans at 12 months differed from the decline over span seen at the two younger ages. The results indicate that the summary scores for this task successfully reflect memory capacity in infancy and that capacity increases over the first year of life. Serial position effects were examined for the two larger spans, namely, those with 3 and 4 items. Repeated measures ANOVAs indicated that novelty scores increased significantly from the first to the last position. These findings, indicative of a clear recency effect, were present at each age for both spans lengths.