Fast development of iconic memory: Infants’ capacity matches adults’

Erik Blaser, PhD. & Zsuzsa Kaldy, PhD.University of Massachusetts Boston

Department of Psychology

MotivationIconic memory is the first visual information buffer, yet

its capacity has never been measured in infants.

GoalDevise a partial report iconic memory paradigm that can be applied both to infants and adults. Compare their capacities.

Iconic memory is the high capacity, short-lived initial store of visual information (Sperling, 1960). Iconic memory has

never been studied in infants.

Here we apply a novel partial report procedure to both infants and adults.

‘Whole report’ tests of ‘very short term memory’ (<300 ms retention) are remarkably low, e.g. just one item for

6.5 month olds (Oakes, Ross-Sheehy, & Luck, 2006).

‘Partial report’ was the key to measuring iconic memory (Sperling, 1960).

Introduction

fixation3.0 seconds

objects are presented1.0 second

a ‘critical pair’ disappears. 500 ms

‘Decoy’ and ‘Target’ reappear2 seconds

presentation partial report cue 2AFC memory test

Red dot shows gaze location,

diameter shows duration

The Target changes color.

The Decoy reappears unchanged

A partial report, 2AFC test of iconic memory for infants and adults

This offset itself acts as a post-cue

Decoy TargetSet Size 8

Rapid developmentRecent fMRI studies in adults attribute iconic memory to persistent activation in higher-order visual areas such as the occipito-temporal cortex, particularly the lateral-occipital complex (LOC; Ruff, Kristjánsson, & Driver, 2007; Wong, Aldcroft, Large, Culham, & Vilis, 2009). Indeed, the occipital lobe is the part of the cortex that matures earliest (Huttenlocher, 1990), and the area corresponding to LOC in adults has been shown to be active at 6.5 months of age (Wilcox et al., 2009).

LO

V1

V1

V2

V2V3

V3

V4

MT+

LO

V1V2V3

V4

V1V2

V3

MT+

Vanish > MaskA

Vanish > MaskLeft Left RightRight

25.0025.00

6.506.50t(8358)t(8358)

-11p < 8.4 x10-11p < 8.4 x10p(Bonf) < 0.000p(Bonf) < 0.000

10.0010.00

5.005.00t(8358)

p < 5.8x10p(Bonf) < 0.009

-7t(8358)

-7p < 5.8x10p(Bonf) < 0.009

Blaser, E., & Kaldy, Z. (2010). Psychological Science, 21, 1643-1645.

Cowan, N. (2001). Behavioral & Brain Sciences, 24, 87-185.

Huttenlocher, P.R. (1990). Neuropsychologia, 28, 517-527.

Oakes, L.M., Ross-Sheehy, S., & Luck, S.J. (2006). Psychological Science,17, 781-7.

Ruff, C.C., Kristjánsson, Á., & Driver, J. (2007). Psychological Science, 18, 901-909.

Sperling, G. (1960). Psychological Monographs: General and Applied, 74, 1-29.

Wilcox, T., Bortfeld, H., Woods, R., Wruck, E., Armstrong, J., & Boas, D. (2009). Neuropsychologia, 47, 657-662.

Wong, Y.J., Aldcroft, A.J., Large, M.E., Culham, J.C., & Vilis, T. (2009). J Neurophys. 102, 3461-3468.

References

This research was supported by National Institutes of Health Grant 2R15EY017985 awarded to the authors

Participants: A total of 62 infants, ~12 per set size, aged 5;00-6;30. 5 ‘Expert’ adults were run, and 12 ‘Naive’/‘Instructed’. Materials: A Tobii T120 eye tracker was used for display and recording. Procedure: Infants were presented with two blocks of 18 test trials. Depending on block, a set of 2, 4, 6, 8, or 10 identically-shaped, but differently-colored items appeared, spaced symmetrically around fixation (adults saw all set sizes). After a 1 sec exposure, a randomly chosen pair of (neighboring) items disappeared. After a 500 ms delay the two items reappeared, with one changed to a new color (Target) and the other unchanged (Decoy). The sudden offset of the pair itself was the partial report post-cue. Since the pair was chosen randomly, any differential treatment of Target vs. Decoy is evidence that information about the items was in memory prior to offset.

Percent correct should be maximal when set size is below capacity, and drop as

capacity is exceeded. This breakpoint is used to estimate capacity (Cowan, 2001).

Our results show nearly identical (5-6 item) iconic memory capacities for 6-month-old infants, instructed adults and expert adults.

Iconic memory, 6 month olds Infant vs. Adult capacity

40

50

60

70

80

90

100

2 4 6 8 1050

55

60

65

70

2 4 6 8 10

Attention grabber (8.5 s)

Presentation (1 s)

Partial report offset post-cue

(500 ms)

Preferential looking memory test (2.5 s)

Fixation (1 s)

Set size

Perc

ent c

orre

ct ***

n.s.

n.s.

**

Preference for Changed object

A trial-by-trial, head-to-head 2AFC comparison

between Target and Decoy. Trials where the Target is

‘preferred’ (fixated longer) are coded as correct.

Capacity estimates shown by red arrows.Expert: 5.75 items; Instructed: 5; Infant: 5; Naive: 2

Set size

Perc

ent c

orre

ct

Preference for Changed object

Blaser & Kaldy, (2010). Psychological Science

Tobii T120

Naive adults had no task knowledge, much

as the infants. Instructed adults were fully

briefed about the experiment and asked to fixate the Changed

item as quickly and for as long as possible. Expert adults were

similarly instructed, and had had previous

experience in visual psychophysics experiments.

Results

We applied a novel partial report test to compare infants’ iconic memory capacity to that of adults.

Our results show a five item iconic memory capacity in 6-month olds.

Infant iconic memory capacity matches adults’ (~5-6 items)(...and, provocatively, exceeds that of naive undergraduates...)

Infants’ and adults’ match in capacity points to particularly rapid development.

Conclusions

InfantsNaive

Expert

Instructed

Replication of set size 4 with independent infant group