P3 response during short-term memory retrieval revisited by a
spatio-temporal analysis
Mehmet Ergen
a,
⁎
, Erol Yildirim
b
, Atilla Uslu
c
, Hakan Gürvit
d
, Tamer Demiralp
c
a
Department of Physiology, School of Medicine, Acibadem University, Maltepe-Gulsuyu 34848, Istanbul, Turkey
b
Department of Psychology, Faculty of Humanities and Social Sciences, Suleyman Sah University, Kartal 34865, Istanbul, Turkey
c
Deparment of Physiology, Istanbul Faculty of Medicine, Istanbul University, Capa 34390, Istanbul, Turkey
d
Deparment of Neurology, Istanbul Faculty of Medicine, Istanbul University, Capa 34390, Istanbul, Turkey
abstractarticle info
Article history:
Received 22 June 2011
Received in revised form 11 December 2011
Accepted 20 February 2012
Available online 28 February 2012
Keywords:
Short-term memory
Memory scanning
Sternberg test
ERP
P3
Memory load
The most reported event related potential (ERP) parameter during short-term memory retrieval has been P3
wave and the association has been built on the relation between P3 latency and reaction times. The aim of
this study is to identify an ERP component that reflects the memory scanning process preceding the decision
making stage which has been associated with the P3 peak. A spatiotemporal analysis was applied on the P3
and pre-P3 period of ERP responses obtained during the retrieval phase of the Sternberg paradigm with two
memory load conditions (3 and 5 letters in the memory set). In the easy task condition with the fastest reac-
tion times (positive probes of 3 letters condition), a single P3 was observed, whereas P3 was split into two
peaks in responses to probe items of more demanding task conditions. The single P3 peak and the later com-
ponents of the split P3 peaks displayed the typical P3 topography. On the other hand, the topographic map-
ping of the earlier peak of the split P3 wave and ascending part of the single P3 peak revealed a right parietal
topography. The onset time of this earlier right lateralized topography was stable among all conditions but it
persisted longer in the high memory load condition. We conclude that the right-lateralized positivity in the
pre-P3 period reflects the memory scanning process followed by the P3 peak with midline parietal topogra-
phy reflecting the decision making process.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
Experimental research on short-term memory function mostly in-
volves tasks that are based on item recognition by asking the partici-
pants to perform a “delayed matching to the sample” task. In these
tasks, participants initially register certain items to be recalled
(encoding phase), after a delay (retention phase) probe items are
presented and participants are asked to judge the probe items as
matching or non-matching with the previously registered items (re-
trieval phase). In the Sternberg paradigm, these three phases are
not overlapping in time which provides a superiority for investigating
one of these phases separately when compared with paradigms such
as n-back.
The original paper by Sternberg reported that reaction times (RTs)
to probes increased linearly with the number of items to be encoded
(memory load), and the author interpreted this memory load depen-
dent delay as the reflection of the memory scanning, which involves
the sequential comparison of the probe to each of the items in the
short-term memory (Sternberg, 1966).
Numerous event-related potential (ERP) studies on Sternberg par-
adigm have concentrated on the P3 response to the probes of the re-
trieval phase, a positive potential that peaks around 300 ms after
stimulus onset (Gomer et al., 1976; Adam and Collins, 1978; Ford
et al., 1979; Donchin, 1981; Pratt et al., 1989; Pelosi et al., 1995;
Starr et al., 1996; Polich, 2007). The common finding of these studies
was the prolongation of the P3 latency due to increasing memory
load, and this had a medium correlation with the of RTs (for a com-
prehensive meta-analysis see Verleger, 1997). This partial correlation
between the P3 latency and the RT has been interpreted as the reflec-
tion of the true prolongation of the memory scanning by Ford et al.
(1979) and Strayer et al. (1987), while the further delay of the RTs
might depend on a more cautious response execution during more
difficult task. However, the distinct electrophysiological signs of this
hypothetical memory scanning process have not yet been described.
A few of the ERP studies on visual Sternberg paradigms (Pratt
et al., 1989; Pelosi et al., 1998; Wolach and Pratt, 2001) reported
the existence of two separate positive peaks in the P3 time window
in response to probe stimuli. In one of these studies (Pratt et al.,
1989), it was reported that the peak latency of the early P3 peak
was relatively constant across memory set sizes as suggestive of an
arousal response resembling the novelty P3a, while the late parietal
P3 was resembling the P3b of the oddball task which increased in
International Journal of Psychophysiology 84 (2012) 205–210
⁎ Corresponding author. Tel.: +90 216 458 08 19; fax: + 90 216 589 84 85.
E-mail addresses: mehmet.ergen@acibadem.edu.tr (M. Ergen), eyildirim@ssu.edu.tr
(E. Yildirim), demiralp@istanbul.edu.tr (T. Demiralp).
0167-8760/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpsycho.2012.02.009
Contents lists available at SciVerse ScienceDirect
International Journal of Psychophysiology
journal homepage: www.elsevier.com/locate/ijpsycho