ArticlePDF Available

Proactive Interference in Human Predictive Learning

Authors:

Abstract and Figures

The impairment in responding to a secondly trained association because of the prior training of another (i.e., proactive interference) is a well-established effect in human and animal research, and it has been demonstrated in many paradigms. However, learning theories have been concerned with proactive interference only when the competing stimuli have been presented in compound at some moment of the training phase. In this experiment we investigated the possibility of proactive interference between elementally-trained stimuli at the acquisition and at the retrieval stages in a behav-ioral task with humans. After training a cue-outcome association we observed retardation in the acquisition of an association between another cue and the same outcome. Moreover, after asymptotic acquisition of the secondly trained association, impairment of retrieval of this secondly trained association was also observed. This finding of proactive interference between elementally-trained cues suggests that interference in predictive learning and other traditional interference effects could be integrated into a common framework.
Content may be subject to copyright.
International Journal of Comparative Psychology, 2002, 15, 55-68.
Copyright 2002 by the International Society for Comparative Psychology
Proactive Interference in Human Predictive Learning
Leyre Castro, Nuria Ortega, and Helena Matute
Universidad de Deusto, Spain
The impairment in responding to a secondly trained association because of the prior training of an-
other (i.e., proactive interference) is a well-established effect in human and animal research, and it
has been demonstrated in many paradigms. However, learning theories have been concerned with
proactive interference only when the competing stimuli have been presented in compound at some
moment of the training phase. In this experiment we investigated the possibility of proactive inter-
ference between elementally-trained stimuli at the acquisition and at the retrieval stages in a behav-
ioral task with humans. After training a cue-outcome association we observed retardation in the
acquisition of an association between another cue and the same outcome. Moreover, after asymptotic
acquisition of the secondly trained association, impairment of retrieval of this secondly trained asso-
ciation was also observed. This finding of proactive interference between elementally-trained cues
suggests that interference in predictive learning and other traditional interference effects could be
integrated into a common framework.
Interference among cues is a central topic in associative learning research.
Cue interference is well represented by Kamin’s early studies (e.g., 1968) with
rats, where he found that the training of two cues in compound after the isolated
training of one of them produced weak responding to the other one. That is, if a
target cue, B, is trained with another cue, A, as a predictor of an outcome (O) after
cue A has been trained as a predictor of that outcome, weak responding to B will
be observed at testing. This is the well known blocking effect. Other cue interfer-
ence effects in the animal learning literature are overshadowing (Pavlov, 1927) and
relative validity (Wagner, Logan, Haberlandt, & Price, 1968). These effects have
been found with humans as well (e.g., Dickinson, Shanks, & Evenden, 1984; Mat-
ute, Arcediano, & Miller, 1996; Price & Yates, 1993; Van Hamme, Kao, &
Wasserman, 1993).
According to associative theories of learning, the acquisition of a cue-
outcome association interferes with the acquisition (e.g., Mackintosh, 1975; Pearce
& Hall, 1980; Rescorla & Wagner, 1972) or the expression (e.g., Miller & Matzel,
1988) of another cue-outcome association only when both cues have been pre-
sented in compound at some point during acquisition (e.g., the blocking effect); if
cues are separately paired with the outcome, none of them will be able to predict
the outcome when the ot her one is present ed, so there should be no interference.
However, recent research with rats (Escobar, Arcediano, & Miller, 2001; Escobar,
Support for this research was provided by Grant PI-2000-12 from Departamento de Educación, Uni-
versidades, e Investigación of the Basque Government to Helena Matute. Leyre Castro was sup-
ported by a F.P.U. fellowship from the Spanish Ministry of Education (Ref. AP99, 14605555) and
Nuria Ortega was supported by a F.P.I. fellowship from the Basque Government (Ref. BFI97.102).
Nuria Ortega is now at the Universidad Nacional de Educación a Distancia, Madrid. We thank Sam
Donaldson, Oskar Pineño, Miguel A. Vadillo and Sonia Vegas for their helpful comments on previ-
ous drafts of this manuscript. Correspondence concerning this article should be addressed to Helena
Matute, Departamento de Psicología, Universidad de Deusto, Apartado 1, 48080 Bilbao, Spain
(hmatute@euskalnet.net).
- 56 -
Matute, & Miller, 2001) and humans (e.g., Matute & Pineño, 1998a, 1998b; Ortega
& Matute, 2000; Pineño, Ortega, & Matute, 2000) has shown that interference
effects can occur between elementally-trained cues, that is, between cues that have
never received compound training. The original finding comes from Matute and
Pineño’s (1998b) studies of predictive learning with humans. They reported that
giving B-O pairings (i.e., a colored light presented with a positive outcome) in
Phase 2 after A-O pairings (i.e., another colored light presented with the same
outcome) in Phase 1 resulted in weak responding to the target cue, A, at test (i.e.,
retroactive interference). They found that this effect was not due to the time
elapsed between Phase 1 and testing, to memory overload, or to blocking by con-
text. Moreover, this effect occurs when the cues that interfere with each other
share the same outcome; that is, there is no interference when the interfering cue,
B, is paired with a different outcome during Phase 2 (Pineño & Matute, 2000;
Escobar, Arcediano, & Miller, 2001).
Although the finding of cue interference between elementally-trained cues
is problematic for contemporary learning theories, it is consistent with results in
the human verbal learning literature. Interference processes have been largely
studied in the paired associate tradition (e.g., see Postman & Underwood, 1973), in
which participants learned the association between a cue item and a response item,
in paradigms such as A-B, A-C or A-B, C-B. More specifically, the reported effect
of interference between elementally-trained cues in predictive learning (e.g., Mat-
ute & Pineño, 1998b), can be seen as an analogue to the retroactive interference
effect observed in the A-B, C-B verbal interference paradigm. In this paradigm,
retroactive interference is said to occur when the learning of C-B after the previous
learning of A-B impairs the subsequent recall of the A-B association (Cheung &
Goulet, 1968; Keppel, Bonge, Strand, & Parker, 1971; Schwartz, 1968).
Moreover, several different manipulations have been shown to have simi-
lar effects in cue interference between elementally-trained cues in predictive learn-
ing and in other retroactive interference paradigms. For example, in Matute and
Pineño’s (1998b) predictive learning experiments, interference was attenuated
when the interfering and target associations were trained in different contexts and
testing occurred in the context in which the target association had been trained.
This is a well demonstrated effect in several other different paradigms: In verbal
learning (Bidoleau & Schlosberg, 1951), in serial learning (Kanak & Stevens,
1992), and in several studies of long-term retention in rats (Chiszar & Spear, 1969;
Spear, 1971; Zentall, 1970). In all of these studies it has been shown that memories
established in a context similar to that of testing are more likely to be retrieved
than memories established in a context different from that of testing. Additionally,
manipulations on the temporal point in which the interpolated learning takes place
have also shown similar results in the interference between elementally-trained
cues in predictive learning as well as in the paired associate tradition. Postman
(1971) reported some experiments in which the first list recall was higher when a
retention interval was introduced just before the test phase, in comparison to the
introduction of a retention interval between the original and the interpolated learn-
ing phases. Moreover, Postman, Stark, and Fraser (1968) showed that when recall
of both the firstly and secondly learned lists was required, the responses from the
secondly learned list were likely to be given first on an immediate test; however,
after a retention interval, the responses from the first list were likely to be given
- 57 -
first. Similarly, Pineño et al. (2000) showed that the introduction of a retention
interval between Phase 2 and testing reduced retroactive interference in predictive
learning, but this did not occur when the retention interval was introduced between
Phases 1 and 2 of training. Seemingly, manipulations that generate a separation
between Phase 2 and testing are able to reduce the activation of the most recently
trained association, therefore allowing the retrieval of the other association.
In sum, this all suggests that interference occurs when the two associations
share a common element and the interfering association is more strongly activated
than the target association at the time of testing (e.g., because of recency or contex-
tual manipulations). Apparently, if the interfering cue is strongly predicting the
outcome when the target cue is presented at testing, retrieval of the outcome repre-
sentation by the target cue is impaired (see Matute & Pineño, 1998a; Pineño et al.,
2000, for further discussion).
A tentative explanation of these effects can be constructed as an extension
of Bouton’s (e.g., 1997) retrieval theory. According to this approach, when two
conflicting associations are trained to asymptote (e.g., when one cue has been
paired with two different outcomes or, extending this theory, when one outcome
has been paired with two different cues), the memory of the second association in
which an event is involved is the most likely to manifest if testing occurs immedi-
ately after training and in the same context. However, when the memory of the
second association is not activated (e.g., because testing takes place in a novel
context or after a period of time), then the firstly trained association is the most
likely to manifest at test. Consequently, the attenuation of the interference effects
would occur not by means of directly activating the target association, but by pre-
venting the interfering association from being strongly activated during testing.
Following this explanation, each of the two associations (regardless of the
order in which they have been trained) should be able to interfere with the other
one as a function of which one is more strongly activated at the time of testing (be
it because of recency or because of other manipulations such as, for example, the
use of contextual or discrete retrieval cues). In this vein, elementally trained asso-
ciations should be able to interfere with each other not only retroactively but also
proactively. Indeed, a recent study of retroactive interference with rats by Escobar,
Matute, and Miller (2001), provides some evidence of proactive interference.
Escobar et al. tested not only for responding to the firstly trained cue, B, but also
for responding to the secondly trained cue, A. They observed that when the sec-
ondly trained association was tested in the context in which the firstly trained asso-
ciation had been trained, weak responding to the secondly trained cue was ob-
tained. This is suggestive of a proactive interference effect. Unfortunately, their
primary concern was retroactive interference, and therefore they did not include
control conditions for proactive interference. Another example of proactive inter-
ference could be the latent inhibition effect, in which the exposure to a stimulus
alone during Phase 1 retards the conditioning to that stimulus during Phase 2.
According to Bouton (1993), the first learning experience (cue-nothing) prevails
over the second one (cue-outcome).
In the present experiment we aimed to further assess the prediction that it
is the relative activation of the associations at testing which produces interference,
by testing for whether elementally trained cues are also sensitive to proactive inter-
ference. With the exception of the preliminary results with rats observed by
- 58 -
Escobar, Matute, and Miller (2001), and the latent inhibition effect in interference
between outcomes, we are not aware of this effect being previously reported in the
predictive learning literature, although proactive interference in other paradigms
with humans (e.g., Bennett, 1975; Brosgole, 1976) as well as with animals (e.g.,
Grant, 1975) is a quite well established effect (but see Crowder, 1967; Kehoe,
1963). This is important in that, if proactive interference between elementally
trained cues were observed to occur in predictive learning, it would add to the
growing body of literature that suggests that interference between cues (e.g., Mat-
ute & Pineño, 1998a) and between outcomes (e.g., Bouton, 1993) in predictive
learning can probably be integrated into a common framewor k along with interfer-
ence effects from other areas.
Overview of the Experiment
We studied proactive interference, trying to answer the following ques-
tions: (1) If a cue is paired to an outcome after a different cue has been trained as a
predictor of the same outcome, will the acquisition of the second cue- outcome
association be retarded?, and (2) once the second cue-outcome association is ac-
quired, could proactive interference still be observed during subsequent retrieval?
During acquisition we assessed how the B-O1 association was acquired
during Phase 2. For three of four groups, Cue A was trained as a predictor of O1
during Phase 1 (Groups SO, Same Outcome). For the other group, Cue A predicted
a different outcome, O2, during Phase 1 (Group DO, Different Outcome). Thus,
we assessed whether the acquisition of an association with one outcome during
Phase 1 in the SO groups impaired their acquisition of the association of B with the
same outcome during Phase 2, as compared to the acquisition in Group DO, that
received the identical exposure to the task but did not acquire an association to O1
during Phase 1. Thus, Groups SO are analogous to the A-B,C-B condition in the
paired associate literature, and the control group, Group DO, is analogous to the A-
B, C-D control condition commonly used in the paired associate paradigm.
In addition, we also assessed proactive interference once the target
association had been acquired. Certainly, it would make no much sense to perform
a direct test of retrieval immediately after all groups have acquired the target
association because all of them would respond appropriately at that time.
However, if our hypothesis is correct, anything that is able to separate Phase 2
from testing (e.g., a retention interval, a contextual change, or even a discrete cue
presented between training and testing) should be able to reduce the activation of
the most recently trained association, consequently allowing the other association
to interfere with it. This idea receives support from studies of retroactive
interference (e. g., Pineño et al., 2000) and can be tested in this experiment by
means of a manipulation that has sometimes been used to vary the activation of
associations in animal experiments concerning interference between outcomes
(e.g., Brooks & Bouton, 1993). The idea is that if a discrete stimulus present
during a certain phase of the study is presented just before testing, during testing
subjects will show behavior that is appropriate to that phase. This manipulation
has also been used within the retroactive interference paradigm with elementally-
trained cues (e.g., Escobar, Matute, & Miller, 2001, with rats; Pineño et al., 2000,
with humans).
- 59 -
Method
Subjects
One hundred and six undergraduate students from Deusto University volunteered for the
study. Random assignment of participants resulted in 30 participants in Group SO, 26 participants in
Group DO, 24 participants in Group SO-Cue2, and 26 participants in Group SO-Novel Cue.
Apparatus
The experiment was run in a large room with personal computers, with the participants re-
sponding through the keyboard. Two sets of fifty and fifty-six participants were run in two different
days. In both days, subjects were randomly distributed across the four experimental groups. Partici-
pants had about 1.5 m of distance between each other, each participant being exposed to a different
experimental condition (and counterbalancing of stimuli) than the two adjacent participants.
Procedure
The design summary is presented in Table 1. In Phase 1, Groups SO, SO-Cue2 and SO-
Novel Cue were exposed to 15 presentations of A, which was always followed by O1, whereas Group
DO was exposed to 15 presentations of A always followed by O2. In addition, during Phase 1 all
groups were exposed to 15 presentations of a retrieval cue, D, intermixed with the presentations of A.
Then, in Phase 2, all groups were exposed to 15 presentations of B, always followed by O1,
interspersed with fifteen presentations of cue C, which was always followed by O3 (i.e., C-O3).
Presentations of C-O3 trials were included in order to prevent cue generalization that would result in
strong responding appropriate to O1 to all cues. Presentations of B and C during Phase 2 were
intermixed with 15 presentations of a retrieval cue, E, in all groups.
Table 1
Design of the Experiment.
Treatment Group
Phase 1 Phase 2 Phase 3
Test
SO AO1, D BO1, CO3, E D B
DO AO2, D BO1, CO3, E D B
SO-Cue2 AO1, D BO1, CO3, E E B
SO-Novel Cue AO1, D BO1, CO3, E F B
Note. A and B were the critical cues. C was included to prevent strong cue generalization. Cues A,
B, and C were blue, red, and green colors, counterbalanced. Presentations of A were followed by O1
in groups SO, SO-Cue2, and SO-Novel Cue, whereas they were followed by O2 in group DO. Pres-
entations of B during trai ning were always followed b y O1 whereas presentations of C were always
followed by O3. O1 was a positive outcome, O2 was a neutral outcome and O3 was a negative out-
come. Cues D, E, and F were used as retrieval cues. These cues were geometric figures (i.e., trian-
gle, circle, and square, counterbalanced) that appeared at a different location from that of the critical
cues and were not associated to any outcome, nor were they a cue to respond (see text).
The other critical difference between groups was located in Phase 3, that is, just before test-
ing. For Groups SO and DO, Phase 3 consisted of one presentati on of cue D, that is, the retrieval cue
that occurred during Phase 1 along with the training of A-O1 (Group SO) or A-O2 (Group DO). This
retrieval cue should be able to prevent the strong activation of the B-O1 association that would other-
wise be expected at the time of testing, because of recency. This weak activation of B at test should
in turn favor the interference effect in Group SO (in which A-O1, an interfering association because it
shared the outcome with the target association, B-O1, could be reactivated) as compared to Group DO
- 60 -
(in which none potentially interfering association had been trained). Moreover, if, according to our
hypothesis, any disruption between Phase 2 and testing could be enough to diminish the strong acti-
vation of the B-O1 association, interference should occur also in Group SO-Novel Cue (for which
Phase 3 consisted of one presentation of a novel cue), as compared to Group DO (for which no inter-
fering association had been trained). Finally, for Group SO-Cue2, Phase 3 consisted of one addi-
tional presentation of the cue included in Phase 2 (i.e., cue E). Because there was no break between
phases, for this group Phase 3 actually consisted of just one more presentation of the cue that was
being presented during Phase 2; thus cue E was used as a control, because just one additional presen-
tation of cue E should be perceived as an additional Phase 2 trial, therefore producing no effect. Trial
order for each phase was pseudorandom, with no more than two trials of the same type occurring in
succession. After Phase 3, all groups were exposed to one test trial with B, which was presented after
the same intertrial interval (ITI) used in the previous phases.
The experimental preparation that we used was the Spy-Radio task, a behavioral task for
use with humans previously used in Pineño et al. (2000) and Pineño & Matute (2000) (A demonstra-
tion version of this program can be downloaded from http://sirio.deusto.es/matute/software.html).
The participant’s task was to rescue a group of refugees by helping them to escape from a war zone in
several trucks. A translation from Spanish of the instructions that the participants received reads as
follows:
Screen 1
Imagine that you are a soldier for the United Nations. Your mission consists of rescu-
ing a group of refugees that are hidden in a ramshackle building. The enemy has detected them
and has sent forces to destroy the building... But, fortunately, they rely on your cunning to escape
the danger zone before that happens.
You have several trucks for rescuing the refugees, and you have to place them in those trucks.
There are two ways of placing people in the trucks:
a) Pressing the space bar repeatedly, so that one person per press is placed in a
truck.
b) Maintaining the space bar pressed down. In such manner, you will be able to
load people very rapidly.
If you rescue a number of persons in a given trip, they will arrive at their destination
alive, and you will be rewarded with a point for each person. You must gain as many points as
possible!
Screen 2
But... your mission will not be as simple as it seems. The enemy knows of your move-
ments and could have placed deadly mines on the road. If the truck hits a mine, it will explode,
and the passengers will die. Each dead passenger will count as one negative point for you.
Fortunately, the colored lights on the SPY-RADIO will indicate to you the state of the
road. The lights can indicate that:
a) The road will be free of mines.
The occupants of the truck will be liberated.
You will gain points.
b) The road will be mined.
The occupants of the truck will die.
You will lose
points.
c) There are no mines, but the road is closed.
The occupants of the truck will nei-
ther die nor be liberated.
You will neither gain nor lose points: You will main-
tain your previous score.
Screen 3
At first, you will not know what each colored light of the SPY-RADIO means. How-
ever, as you gain experience with them, you will learn to interpret what they mean.
Thus, we recommend that you:
a) Place more people in the truck the more certain you are that the road will be free
of mines (keep the space bar continuously pressed down ONLY if you are com-
pletely sure that there are no mines, because in this way you will put a lot of peo-
ple in the truck...).
- 61 -
b) Introduce fewer people in the truck the more certain you are that the road is
mined.
The fourth instruction screen informed participants that they would possibly rescue refu-
gees in different towns; these towns would provide different contexts that could be manipulated by
the experimenter. Although contextual changes were not used in the present experiment, we main-
tained the four instructional screens of this program in order to avoid making more changes than
necessary between different series of experiments using this same preparation. A translation of the
fourth screen reads as follows:
Screen 4
Finally, it is important to know that your mission may take place in several different
towns. The colors on the SPY-RADIO can mean the same or a very different thing depending on the
town in which you are. Thus, it is important to pay attention to the message that indicates the place
in which you are. If you travel to another town, the message indicating the name of the town will
change. When a change of destination is occurring, you will read the message “Traveling to an-
other town”, so you will be continuously informed about such changes. Nevertheless, sometimes you
might end up returning to the same town even if you have seen the message that indicates that you
are traveling. Do not worry if all this looks like very complex at this point. Before we start, you will
have the opportunity to see the location of everything (radio, town name, messages, scores, etc.) on
the screen, and to ask the experimenter about anything that is unclear.
The top of the screen showed a “spy-radio” that consisted of a panel in which six colored
lights could be presented. Cues A, B, and C were blue, red, and green lights in the spy radio, coun-
terbalanced. In this experiment, each time that a cue was presented, all six panels were illuminated
with the color of that cue. Cue duration was 3 s. During the ITIs, the lights were turned off (i.e.,
gray). The ITI duration was random with a range between 3 and 7 s, and a mean of 5 s. While the
lights were on, each response (i.e., pressing the space bar once) placed one refugee in the truck. If
the participant maintained the space bar pressed down while the lights were on, up to 30 refugees per
second could be placed in the truck with the computers used in this experiment. On each trial, the
termination of the cue was immediately followed by the onset of the outcome.
Outcome 1 (O1) consisted of (a) the message “[N] refugees safe at home!!!” (with [N] be-
ing the number of refugees introduced in the truck during the cue presentation) and, (b) gaining one
point for each refugee who was liberated. Outcome 2 (O2) consisted of (a) the message “Road
closed” and, (b) maintaining previous score. Outcome 3 (O3) consisted of (a) the message “[N] refu-
gees have been killed!!!” and, (b) losing one point for each refugee who died in the truck. Outcome
messages were presented for 3 s. Outcomes were not counterbalanced because our dependent vari-
able was the mean number of responses given during the presentation of B. Thus, B had to be always
paired to an appetitive outcome. Otherwise, strong or weak responding at test would depend on
outcome counterbalancing rather than on whether interference exists or not.
The number of refugees that participants risked taking in each truck was our dependent
variable. Presumably, the more certain they were that the trip would be successful (i. e., O1), the
greater number of refugees they would take, whereas the more certain they were that the truck would
explode (i.e., O3), the smaller number of refugees they would take. Additionally, because introducing
refugees did not have any effect on the score when O2 followed the cue, we expected the participants
to take no refugees when the cue was followed by O2.
One score panel on the screen provided information during the experiment. The panel
showed the number of people that the participant was introducing in the truck on each trial. Although
bar presses that occurred while the outcome message was present had no consequences, the score
panel remained vi sible during the presentation of the outcome and showed the number of people that
had been boarded while the cue was present. At the termination of the outcome, this panel was set to
zero. Responses that occurred during the ITIs had no consequences and were not reflected in the
panel. Thus, only responses that occurred while a cue was presented resulted in refugees traveling in
the truck and participants gaining or losing points (if O1 or O3 followed the cue) or keeping the previ-
ous score (if O2 followed the cue).
The ret rieval cues (i.e., cues D, E, and F) were geometrical figures (i.e., a cir cle, a triangle,
and a square, counterbalanced) that appeared at the left-bottom corner of the screen rather than in the
spy-radio. Thus, these cues were not associated with the possibility to introduce people in the truck,
- 62 -
nor with any outcome. They simply appeared in the screen subject to the identical ITIs as the other
cues; therefore, they never appeared at the same time as any of the regular cues but an interval be-
tween 3 and 7 s occurred between these types of events. The instructions did not give any informa-
tion on the existence of these retrieval cues. Moreover, if any response was given during the presen-
tation of these cues, it was treated in the same way as responses that occurred during the ITIs, that is,
they had no consequences (and were not reflected in the score panel, nor in the panel that showed the
number of people introduced in the truck). The duration of these cues was identical to that of the
other cues (i.e., 3 s). The different phases of the experiment, including the test phase, were presented
without interruption.
Preanalysis Treatment of the Data
We normally use a data selection criterion in order to ensure that participants are paying at-
tention to the experiment and have acquired the discrimination during the phase in which this dis-
crimination occurs (i.e., Phase 2 in the present experiment). According to this criterion, in the present
experiment, responding to B during the last trial in which it is presented during training has to be
higher than responding to the last trial of C (e.g., Ortega & Matute, 2000; Pineño et al., 2000). Fol-
lowing this criterion, the data from 1 participant from Group DO and 1 participant from Group SO-
Cue2 were eliminated from the analyses. An alpha level of .05 was adopted for tests of statistical
significance.
Results
Interference during Training
The results of the two training phases are shown in Figure 1. As can be
seen in this figure (left panel) training in Phase 1 proceeded smoothly, showing
that all four groups had a good learning level of the first cue-outcome association
at the end of that phase: a high number of responses in Groups SO, SO-Cue2 and
SO-Novel Cue and a low number of r esponses in Group DO, as was expected du e
to the different outcomes that they received.
Most important, groups that received O1 during Phase 1 showed proactive
interference, as they were slower than the group receiving O2 during Phase 1
(Group DO) in the acquisition of the B-O1 association during Phase 2 (see Figure
1, right panel). A 4 (Group) x 5 (Blocks of trials) analysis of variance (ANOVA)
on responding to B during Phase 2 confirmed these impressions. This ANOVA
revealed a main effect of Group, F(3, 100) = 4.42, a main effect of blocks of trials,
F(4, 400) = 139.95, as well as an interaction, F(12, 400) = 1.10. Planned compari-
sons revealed differences on the first block of training trials with B between Group
DO and all the groups that received training with the same outcome in Phase 1:
with Group SO, F(1, 100) = 9.18, with Group SO-Cue 2, F(1, 100) = 11.41, and
with Group SO-Novel Cue, F(1, 100) = 3.54 (p =.06 in this case). Differences
were also significant in the second block of B trials between Group DO and Group
SO, F(1, 100) = 6.24, Group SO-Cue 2, F(1, 100) = 8.46, Group SO-Novel Cue,
F(1, 100) = 3.96. This pattern fluctuated thereafter. These results show a proactive
interference effect in the acquisition of an association that shares the outcome with
a previously trained association.
- 63 -
0
10
20
30
40
50
60
70
12345 12345
Training Block Trials
Mean Number of Responses
SO
DO
SO-Cue2
SO-Novel Cue
Phase 1 Phase 2
A -> O1
A -> O2
C -> O3
B -> O1
Figure 1. Mean number of responses to A, B and C during training. Error bars represent standard
error of the means.
Interference at Test
The next step was to assess interference at the retrieval stage. Figure 2
shows the mean number of responses during testing. As can be seen in this figure,
proactive interference was again observed at testing in those groups in which the
target association was no more strongly activated than the interfering association at
the time of testing. Groups SO and SO-Novel Cue showed weak responding to B
at test, compared to Groups DO (in which no interfering association had been
trained) and SO-Cue2 (for which nothing potentially disruptive had been presented
between the training of the target association and the test phase, which allowed the
target association to be strongly activated during testing, due to recency).
In order to confirm these impressions, we performed an analysis of covari-
ance (ANCOVA) on the test of B, including the last block of trials of Phase 2 as a
covariate to control for the differential baseline levels of responding at the end of
Phase 2 training. This ANCOVA revealed a main effect of group, F(3, 99) = 4.95 ,
and planned comparisons showed that responding in Group SO was weak as com-
pared to Group DO, F(1, 99) = 3.12, and to Group SO-Cue2, F(1, 99) = 5.04. This
weak responding in Group SO as compared to Group DO suggests an interference
effect due to the presentation of the retrieval cue from Phase 1 before testing. In
principle, this can occur either through direct activation of the interfering associa-
tion by the retrieval cue, or through a reduction in the activation of the recent B-O
association. In Group DO, however, in which none potentially interfering associa-
tion had been trained, the target association could be expressed. Additionally,
responding to B remains strong in Group SO-Cue2 because, before testing, these
participants were just given one additional presentation of the Phase 2 cue that was
- 64 -
being trained along with B. Because there is no break between phases, for Group
SO-Cue2 the presentation of the Phase 2 cue is actually one more trial of Phase 2,
which allows for a strong activation of the B-O1 association at the time of testing
(because of recency) and, consequently, strong responding to B is observed.
0
10
20
30
40
50
60
70
SO DO SO-Cue2 SO-Novel Cue
Group
Mean Number of
Responses
Figure 2. Mean number of responses to B during the test trial. Error bars represent standard error of
the means.
Group SO-Novel Cue also showed weak responding to B as compared to
Group DO, F(1, 99) = 8.54, and to Group SO-Cue2, F(1, 99) = 11.43. This sug-
gests that a retrieval cue for the interfering association (e.g., as in Group SO) is not
a necessary condition to observe this effect. Instead, the interference effect can be
obtained when any disruption between Phase 2 and testing occurs. This disruption
reduces the strong activation of the recent B-O1 association, allowing for the reac-
tivation of the interfering association, A-O1.
However, it might be possible that the differences found at testing oc-
curred not due to a decrement in responding in Groups SO and SO-Novel Cue, but
to an increment in Groups DO and SO-Cue 2. To assess this alternative interpreta-
tion, we compared the responding at the last block of trials of Phase 2 with repond-
ing at testing. A 4 (Group) x 2 (Trials: Last Block vs. Test) ANOVA revealed a
main effect of Group, F(3,100) = 5.96, a main effect of Trial, F(1,100) = 15.87, as
well as an interaction, F(3,100) = 4.07. Planned comparisons revealed differences
between the end of Phase 2 and the test trial in Group SO, F(1,100) = 8.48 (M =
59.28, SEM = 1.42, and M = 47.86, SEM = 4.66, respectively) and in Group SO-
Novel Cue, F(1,100) = 20.07, (M = 59.67, SEM = 2.68, and M = 40.80, SEM =
5.34, respectively). On the other hand, differences between the end of Phase 2 and
the test trial were not significant in Group DO, F(1,100) = 1.26 (M = 65.50, SEM =
1.30, and M = 60.68, SEM = 3.00, respectively) or in Group SO-Cue 2, F(1,100) =
0.09 (M = 59.55, SEM = 1.44, and M = 60.95, SEM = 1.77, respectively). Thus,
thes e analyses show that the observed r esults are a genuine interference effect du e
to a decrement in responding in Groups SO and SO-Novel Cue.
These results support the hypothesis that, like retroactive interference, pro-
active interference also occurs when the two associations share a common element
and the interfering association is more strongly activated than the target association
at the time of testing. In the present case, this was achieved through the presenta-
tion of different retrieval cues. Moreover, the interference effect occurred not by
means of directly activating the interfering association, but by preventing the target
- 65 -
association from being strongly activated and therefore allowing the interfering
association to interfere with it at the time of testing.
Discussion
The results of this experiment showed proactive interference at training as
well as proactive interference at testing between elementally-trained cues. That is,
during both the acquisition stage and the retrieval stage, learning of the firstly
trained association can impair responding to the secondly trained association. As
we expected, when two associations that share the same outcome are trained, each
of both associations is able to interfere with the other one retroactively (as the stud-
ies reported in the Introduction showed) and proactively as well, as the present
results show.
The interference effect found here results from a failure to retrieve the in-
formation concerning the target association (i.e., B-O1) as a consequence of the
previous training of another association that shared the same outcome (i.e., A-O1).
Matute and Pineño (1998a, 1998b) suggested that a cue-outcome association can
retroactively interfere with the target association when the two of them share the
same outcome and the interfering association is more strongly activated at test.
This stronger activation of the interfering association impairs the retrieval of the
target association, and consequently weakens responding. Applied to the present
experiment on proactive interference, a similar process is observed: Acquisition is
slower because the previously trained association (A-O1) is still active in memory
during those early trials of training in which the target association (B-O1) is being
trained. Then, once the two associations have been acquir ed, proactive interfer-
ence can again be observed through the activation mechanism at the postacquisi-
tion stage if appropriate manipulations are conducted. As has been observed in
this experiment, if testing occurs immediately after training on B, good responding
to B is observed (i.e., Group SO-Cue2). However, presenting before the test of B a
retrieval cue that occurred along with the training of the interfering association, or
a novel cue, impairs the retrieval of the secondly trained association (presumably
not by directly activating the interfering association, but by preventing the sec-
ondly trained association from being strongly activated at testing).
Proactive interference, as well as retroactive interference between elemen-
tally-trained cues, cannot be predicted by traditional learning theories because
these theories assume that compound training of cues is a necessary condition for
interference to occur either during acquisition (e.g., Rescorla & Wager, 1972) or
during retrieval (e.g., Miller & Matzel, 1988). However, results on proactive and
retroactive interference in predictive learning are consistent with those reported in
other interference paradigms. In the paired associate tradition, retroactive and
proactive interference could be assessed with either the A-B, C-B or the A-B, A-C
paradigms, in which retroactive interference refers to weak recall of the firstly
trained association and proactive interference refers to weak acquisition or recall of
the secondly trained association (e.g., Bäuml, 1996; Bennett, 1975; Bidoleau &
Schlosberg, 1951; Brosgole, 1976; Chandler, 1993; Chandler & Gargano, 1998;
Postman, 1971; Postman et al., 1968; Underwood, 1966). The A-B, C-B paradigm
parallels interference between elementally-trained cues in predictive learning, both
retroactively (Escobar, Matute, & Miller, 2001; Matute & Pineño, 1998a, 1998b;
- 66 -
Ortega & Matute, 2000; Pineño et al., 2000) and proactively (present experiment).
On the other hand, interference effects in the A-B, A-C paradigm can be seen as
analogous to interference effects between elementally trained outcomes, such as
extinction and counterconditioning (e.g., Bouton, 1993; Pineño & Matute, 2000).
Moreover, recent research has shown that interference between compounded ele-
ments does not only occur between cues trained in compound (e.g., blocking;
Kamin, 1968), but also between outcomes trained in compound (e.g., Esmorís-
Arranz, Miller & Matute, 1997; Miller & Matute, 1998). In summary, retroactive
and proactive interference, both between compound or elemental cues and out-
comes, can be observed in a variety of paradigms, and their common characteris-
tics suggest the viability of a common explanation.
The present results, along with those on retroactive interference (e.g.,
Escobar, Matute, & Miller, 2001; Matute & Pineño, 1998b; Pineño et al., 2000)
suggest that (1) in the absence of retrieval or contextual cues the most recently
trained association can interfere with the expression of the firstly trained associa-
tion, and (2) if contextual, retrieval, or even novel cues disrupt the effect of re-
cency, the firstly trained association tends to prevail and can impair retrieval of the
most recent association (conversely, in the case of retroactive interference, the
interference effect is attenuated, rather than increased, when the test phase is sepa-
rated from the training phase; see Pineño et al., 2000).
A similar mechanism for interference has also been proposed by Chandler
and Gargano (1998) in relation to forced-choice recognition tests. They suggested
that retroactive interference effects occur because, by default, “recent items are
stronger and therefore more likely to be sampled at the expense of the target” (p.
227). The present results are consistent with this view and extend it by showing
that older items, rather than recent items, tend to prevail when there is something
(e.g., a novel, nontarget cue) that disrupts the continuum between training and
testing.
This hypothesis of the strongly activated association can be seen as an ex-
tension of Bouton´s theory (1993, 1994) concerning interference between out-
comes. In studies of interference between outcomes, two associations are also
trained in different phases. However, instead of training two different cues with
the same outcome, a single cue is paired with two different outcomes in different
phases (e.g., extinction and counterconditioning). That is, a cue is paired with one
outcome in Phase 1 (i.e., A-O1), and with no outcome or with a different outcome
in Phase 2 (i.e., A-no O or A-O2). In this case, behavior appropriate to Phase 2 is
normally observed at testing, due to recency (i.e., retroactive interference). How-
ever, if the effects of recency are overcome (e.g., through a retention interval in-
troduced before testing), behavior appropriate to Phase 1 is observed (i.e., sponta-
neous recovery, Pavlov, 1927; Bouton, 1993). Moreover, if a nontarget cue that
was present during a certain phase of the study is presented before testing, subjects
will show behavior during testing that is appropriate to that phase (Brooks, 2000;
Brooks & Bouton, 1993). Finally, if we take into account that the memory of the
first association in which an event is involved is the most likely to manifest when
the memory of the second one is not strongly activated (e.g., as in spontaneous
recovery; Bouton, 1997), it could be argued that interference effects occur not by
the direct activation of the interfering association, but by the prevention of the
- 67 -
target association from being strongly activated, which therefore allows the inter-
fering association to interfere with it at testing.
The present results add support to the idea that this approach can be ap-
plied to both retroactive and proactive interference, both between cues and be-
tween outcomes (e.g., Matute & Pineño, 1998a). Indeed, interference between
elementally-trained cues and interference between elementally-trained outcomes
can be seen as symmetrical effects: Two different cues are paired with the same
outcome or one cue is paired with two different outcomes. Further testing of the
particular account that we have provided here will decide about its adequacy.
References
Bäuml, K. H. (1996). Revisiting an old issue: Retroactive interference as a function of the
degree of original and interpolated learning. Psychonomic Bulletin and Review, 3, 380-384.
Bennett, R. W. (1975). Proactive Interference in Short-Term Memory: Fundamental Forget-
ting Processes. Journal of Verbal Learning and Verbal Behavior, 14, 123-144.
Bidoleau, I. & Scholsberg, H. (1951). Similarity in stimulating conditions as a variable in
retroactive inhibi tion. Journal of Experimental Psychology, 41, 199-204.
Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of
Pavlovian learning. Psychological Bulletin, 114, 80-99.
Bouton, M. E. (1994). Context, ambiguity, and classical conditioning. Current Directions in
Psychological Science, 3, 49-53.
Bouton, M. E. (1997). Signals for whether versus when an event will occur. In M. E. Bouton
& M. S. Fanselow (Eds.), Learning, motivation, and cognition: The functional behaviorism of Robert
C. Bolles (pp. 385-409). Washington, DC: American Psychological Association.
Brooks, D. C. (2000). Recent and remote extincti on cues reduce spontaneous recovery. The
Quarterly Journal of Experimental Psychology, 53B, 25-58.
Brooks, D. C., & Bouton, M. E. (1993). A retrieval cue for extinction attenuates spontaneous
recovery. Journal of Experimental Psychology: Animal Behavior Processes, 19, 77-89.
Brosgol e, L. (1976). Phenomenal determination of retroactive and proactive interference.
Psychological Reports, 38, 1093-1094.
Chandler, C. C. (1993). Accessing related events increases retroactive interference in a
matching recognition test. Journal of Experimental Psychology: Learning, Memory and Cognition,
19, 967-974.
Chandler, C. C., & Gargano, G. J. (1998). Retr ieval processes that produce interference in
modified forced-choice recognition tests. Memory and Cognition, 26, 220-231.
Cheung, C. G., & Goulet, L. R. (1968). Retroactive inhibition of R-S associations in the A-
B, B-C, C-B paradigms. Journal of Experimental Psychology, 76, 327-328.
Chiszar, D. A., & Spear, N. E. (1969). Stimulus change, reversal learning, and retention in
the rat. Journal of Comparative and Physiological Psychology, 69, 190-195.
Crowder, R. G. (1967). Proactive and retroactive inhibition in the retention of a T-maze habit
in rats. Journal of Experimental Psychology, 74, 167-171.
Dickinson, A., Shanks, D., & Evenden, J. (1984). Judgement of act-outcome contingency:
The role of selective attribution. Quarterly Journal of Experimental Psychology, 36A, 29-50.
Escobar, M., Arcediano, F., & Miller, R. R. (2001). Conditions favoring retroactive interfer-
ence between antecedent events (cue competition) and between subsequent events (outcome competi-
tion). Psychological Bulletin and Review, 8, 691-697.
Escobar, M., Matute, H., & Miller, R. R. (2001). Cue trained apart compete for behavioral
control in rats: Convergence with the associative interference literature. Journal of Experimental
Psychology: General, 130, 97-115.
Esmorís-Arranz, F. J., Miller, R. R., & Matute, H. (1997). Blocking of antecedent and sub-
sequent events. Journal of Experimental Psychology: Animal Behavior Processes, 23, 145-156.
Grant, D. S. (1975). Proactive Interference in pigeon short-term memory. Journal of Ex-
perimental Psychology: Animal Behavior Processes, 104, 207-220.
Kamin, L. J. (1968). "Attention-like" processes in classical conditioning. In M.R. Jones
(Ed.): Miami symposium on the prediction of behavior: Aversive stimulation (pp 9-31). Miami, FL:
University of Miami Press.
- 68 -
Kanak, N. J., & Stevens, R. (1992). PI and RI in serial learning as a function of environ-
mental context. Applied Cognitive Psychology, 6, 589-606.
Kehoe, E. J. (1963). Effects of prior and interpolated learning on retention in pigeons. Jour-
nal of Experimental Psychology, 65, 537-545.
Keppel, G., Bonge, D., Strand, B. Z., & Parker, J. (1971). Direct and indirect interference in
the recall of paired associates. Journal of Experimental Psychology, 88, 414-422.
Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli
with reinforcement. Psychological Review, 82, 276-298.
Matute, H., Arcediano, F., & Miller, R. R. (1996). Test Question modulates cue competition
between causes and between effects. Journal of Experimental Psychology: Learning, Memory and
Cognition, 22, 182-196.
Matute, H., & Pineño, O. (1998a). Cue Competition in the absence of compound training:
Its relati on to paradigms of competition between outcomes. In D. L. Medin (Ed.). The psychology of
learning and motivation, Vol. 38, (pp. 45-81). San Diego, CA: Academic Press.
Matute, H., & Pineño, O. (1998b). Stimulus competition in the absence of compound condi-
tioning. Animal Learning and Behavior, 26, 3-14.
Miller, R. R., & Matute, H. (1998). Competition between outcomes. Psychological Science,
9, 146-149.
Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: A response rule for the
expression of associations. In G. H. Bower (Ed.), The psychology of learning and motivation, Vol. 22,
(pp. 51-92). San Diego, CA: Academic Press.
Ortega, N., & Matute, H. (2000). Interference between elementally-trained stimuli can take
place in one trial. Learning and Motivation, 31, 323-344.
Pavlov, I. P. (1927). Conditioned reflexes. London: Clarendon Press.
Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: Variations in the effec-
tiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532-552.
Pineño, O., & Matute, H. (2000). Interference in human predictive learning when associa-
tions share a common element. International Journal of Comparative Psychology, 13, 16-33.
Pineño, O., Ortega, N., & Matute, H. (2000). The relative activation of the associations
modulates interference between elementally-trained cues. Learning and Motivation, 31, 128-152.
Postman, L. (1971). Transfer, interference and forgetting. In K. W. Kling & L. A. Riggs
(Eds.), Experimental Psychology (3rd ed., pp.1019-1132). New York: Holt, Rinehart and Winston.
Postman, L., Stark, K., & Fraser, J. (1968). Temporal changes in interference. Journal of
Verbal Learning and Verbal Behavior, 7, 672-694.
Postman, L., & Underwood, B. J. (1973). Critical issues in interference theory. Memory and
Cognition, 1, 19-40.
Price, P.C., & Yates, J.F. (1993). Judgmental overshadowing: Further evidence of cue inter-
action in contingency judgment. Memory and Cognition, 21, 561-572.
Rescor la, R. A., & Wagner, A. R. (1972). A theor y of Pavlovian conditioning: Variations in
the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.),
Classical conditioning II: Current research and theory (pp. 64-99). New York: Appleton.
Schwartz, M. (1968). Effect of stimulus class on transfer and RI in the A-B, A-C paradigm.
Journal of Verbal Learning and Verbal Behavior, 7, 189-195.
Spear, N. E. (1971). Forgetting as retrieval failure. In W. K. Honig. & P. H. R. James (Eds.),
Animal memory (pp. 45-109). New York: Academic Press.
Underwood, B. J. (1966). Experimental psychology (2nd ed.). New York: Appleton-Century-
Crofts.
Van Hamme, L. J., Kao, S-F., & Wasserman, E. A. (1993). Judging int erevent relati ons:
From cause to effect and from effect to cause. Memory and Cognition, 21, 802-808.
Wagner, A. R., Logan, F. A., Haberlandt, K., & Price, T. (1968). Stimulus selection and a
“modified continuity theory”. Journal of Experimental Psychology, 76, 171-180.
Zentall, T. R. (1970). Effects of context change on forgetting in rats. Journal of Experimen-
tal Psychology, 86, 440-448.
Received July 18, 2002.
Revision received January 27, 2003.
Accepted January 28, 2003.
... With regards to semantic proactive interference, the results show a modest ICC with no relevant change between the two assessments, which can be interpreted as a quite stable measure. This is consistently described in associative learning research [52]. The effect of contextual cues in acquisition and/or retrieval tends to disrupt the recency effect, that is, the firstly trained association (list 1) tends to prevail affecting the retrieval of the most recent one (list 2). ...
Article
Full-text available
Background: Episodic memory testing is fundamental for the diagnosis of Alzheimer's disease (AD). Although the Free and Cued Selective Reminding Test (FCSRT) is widely used for this purpose, it may not be sensitive enough for early detection of subtle decline in preclinical AD. The Memory Binding Test (MBT) intends to overcome this limitation. Objectives: To analyze the test-retest reliability of the MBT and its convergent validity with the FCRST. Methods: 36 cognitively healthy participants of the ALFA Study, aged 45 to 65, were included for the test-retest study and 69 for the convergent analysis. They were visited twice in a period of 6 ± 2 weeks. Test-retest reliability was determined by the calculation of the intra-class correlation coefficient (ICC). Score differences were studied by computing the mean percentage of score variation between visits and visualized by Bland-Altman plots. Convergent validity was determined by Pearson's correlations. Results: ICC values in the test-retest reliability analysis of the MBT direct scores ranged from 0.64 to 0.76. Subjects showed consistent practice effects, with mean amounts of score increasing between 10% and 26% . Pearson correlation between MBT and FCSRT direct scores showed r values between 0.40 and 0.53. The FCSRT displayed ceiling effects not observed in the MBT. Conclusions: The MBT shows adequate test-retest reliability and overall moderate convergent validity with the FCSRT. Unlike the FCSRT, the MBT does not have ceiling effects and it may therefore be especially useful in longitudinal studies, facilitating the measurement of subtle memory performance decline and the detection of very early AD.
... En los últimos años se ha venido acumulando una evidencia empírica importante a favor de que determinadas teorías asociativas, surgidas originariamente en el campo del condicionamiento animal para explicar estos fenómenos de interferencia, ofrecen una buena explicación de fenómenos de interferencia equivalentes en el aprendizaje causal humano (e.g., la teoría de la recuperación de Bouton, 1993). En particular, la evidencia se ha venido acumulando en el terreno de los fenómenos tanto de interferencia entre diferentes resultados de una misma señal, como de interferencia entre diferentes señales de un mismo resultado (Castro, Ortega y Matute, 2002;Matute y Pineño, 1998a, 1998bPineño, Ortega y Matute, 2000;Vila y Rosas, 2001a, b). ...
Book
Full-text available
Este libro nace a partir de un symposium titulado "Extinción y recuperación de la información en aprendizaje causal: perspectivas teóricas" organizado en el marco del XV Congreso de la Sociedad Española de Psicología Comparada, celebrado en Barcelona en septiembre del año 2003 y en el que la mayoría de los grupos hispanos dedicados a este tema presentaron los últimos avances de sus investigaciones y reflexiones teóricas.
... En los últimos años se ha venido acumulando una evidencia empírica importante a favor de que determinadas teorías asociativas, surgidas originariamente en el campo del condicionamiento animal para explicar estos fenómenos de interferencia, ofrecen una buena explicación de fenómenos de interferencia equivalentes en el aprendizaje causal humano (e.g., la teoría de la recuperación de Bouton, 1993). En particular, la evidencia se ha venido acumulando en el terreno de los fenómenos tanto de interferencia entre diferentes resultados de una misma señal, como de interferencia entre diferentes señales de un mismo resultado (Castro, Ortega y Matute, 2002;Matute y Pineño, 1998a, 1998bPineño, Ortega y Matute, 2000;Vila y Rosas, 2001a, b). ...
... En los últimos años se ha venido acumulando una evidencia empírica importante a favor de que determinadas teorías asociativas, surgidas originariamente en el campo del condicionamiento animal para explicar estos fenómenos de interferencia, ofrecen una buena explicación de fenómenos de interferencia equivalentes en el aprendizaje causal humano (e.g., la teoría de la recuperación de Bouton, 1993). En particular, la evidencia se ha venido acumulando en el terreno de los fenómenos tanto de interferencia entre diferentes resultados de una misma señal, como de interferencia entre diferentes señales de un mismo resultado (Castro, Ortega y Matute, 2002;Matute y Pineño, 1998a, 1998bPineño, Ortega y Matute, 2000;Vila y Rosas, 2001a, b). ...
Conference Paper
Full-text available
Trabajos en extenso presentados en el Symposium titulado "Extinción y recuperación de la información en aprendizaje causal: perspectivas teóricas" organizado en el marco del XV Congreso de la Sociedad Española de Psicología Comparada, celebrado en Barcelona en septiembre del año 2003 y en el que la mayoría de los grupos hispanos dedicados a este tema presentaron los últimos avances de sus investigaciones y reflexiones teóricas.
... En los últimos años se ha venido acumulando una evidencia empírica importante a favor de que determinadas teorías asociativas, surgidas originariamente en el campo del condicionamiento animal para explicar estos fenómenos de interferencia, ofrecen una buena explicación de fenómenos de interferencia equivalentes en el aprendizaje causal humano (e.g., la teoría de la recuperación de Bouton, 1993). En particular, la evidencia se ha venido acumulando en el terreno de los fenómenos tanto de interferencia entre diferentes resultados de una misma señal, como de interferencia entre diferentes señales de un mismo resultado (Castro, Ortega y Matute, 2002;Matute y Pineño, 1998a, 1998bPineño, Ortega y Matute, 2000;Vila y Rosas, 2001a, b). ...
... However, these differences disappeared during the last block of trials, t(159) = 1.26, p = .21. This poor performance in response to A was a natural consequence of the previous pairings of cue A with a different outcome in Phase 1; it is perfectly consistent with previous demonstrations of proactive interference in memory experiments (Keppel & Underwood, 1962) and similar associative learning studies (Amundson, Escobar, & Miller, 2003; Castro, Ortega, & Matute, 2002). ...
Article
Current associative theories of contingency learning assume that inhibitory learning plays a part in the interference between outcomes. However, it is unclear whether this inhibitory learning results in the inhibition of the outcome representation or whether it simply counteracts previous excitatory learning so that the outcome representation is neither activated nor inhibited. Additionally, these models tend to conceptualize inhibition as a relatively transient and cue-dependent state. However, research on retrieval-induced forgetting suggests that the inhibition of representations is a real process that can be relatively independent of the retrieval cue used to access the inhibited information. Consistent with this alternative view, we found that interference between outcomes reduces the retrievability of the target outcome even when the outcome is associated with a novel (non-inhibitory) cue. This result has important theoretical implications for associative models of interference and shows that the empirical facts and theories developed in studies of retrieval-induced forgetting might be relevant in contingency learning and vice versa.
... Apart from our learning task, there are only three tasks that have been used to obtain interference between cues, even though with different frequency and success. The most frequently and successfully used is the spy-radio task (Castro, Ortega, & Matute, 2002;Escobar et al., 2002;. In this task, participants have to rescue as many refugees as possible in a war zone plagued with hidden mines by repeatedly pressing the space bar to place the refugees in a series of trucks. ...
Article
In an interference-between-cues design, the expression of a learned Cue A --> Outcome 1 association has been shown to be impaired if another cue, B, is separately paired with the same outcome in a second learning phase. In the present study, we assessed whether this interference effect is mediated by participants' previous causal knowledge. This was achieved by having participants learn in a diagnostic situation in Experiment 1a, and then by manipulating the causal order of the learning task in Experiments 1b and 2. If participants use their previous causal knowledge during the learning process, interference should only be observed in the diagnostic situation because only there we have a common cause (Outcome 1) of two disjoint effects, namely cues A and B. Consistent with this prediction, interference between cues was only found in Experiment 1a and in the diagnostic conditions of Experiments 1b and 2.
Article
Full-text available
In this article we describe some of the experimental software we have developed for the study of associative human learning and memory. All these programs have the appearance of very simple video games. Some of them use the participants' behavioral responses to certain stimuli during the game as a dependent variable for measuring their learning of the target cue-outcome associations. Some others explicitly ask participants to rate the degree of relationship they perceive between the cues and the outcomes. These programs are implemented in Web pages using JavaScript, which allows their use both in traditional laboratory experiments as well as in Internet-based experiments.
Article
Simultaneous acquisition of two serial lists produced retroactive and proactive interference, so long as the lists were organized as one preceding the other. Such an organization was accomplished in a variety of circumstances, as in the case of placing one above the other with an item of a different color separating the lists. In this instance, the lists were clearly organized into top and bottom, with the top half coming before the bottom. Recall for the two tasks was assessed 10 min. and 24 hr. after original learning. The recall of the list perceived as coming last was significantly superior after a 10-min. delay (retroactive interference) but deteriorated with time so that the two equalized after 24 hr. (proactive interference). The fact that retroactive and proactive interference were produced in the absence of a temporal separation between original and interpolated learning indicates that these phenomena are probably the product of a dynamic interaction in memory rather than extinction occurring at the point of interpolation.
Chapter
This chapter describes the potential explanatory power of a specific response rule and its implications for models of acquisition. This response rule is called the “comparator hypothesis.” It was originally inspired by Rescorla's contingency theory. Rescorla noted that if the number and frequency of conditioned stimulus–unconditioned stimulus (CS–US) pairings are held constant, unsignaled presentations of the US during training attenuate conditioned responding. This observation complemented the long recognized fact that the delivery of nonreinforced presentations of the CS during training also attenuates conditioned responding. The symmetry of the two findings prompted Rescorla to propose that during training, subjects inferred both the probability of the US in the presence of the CS and the probability of the US in the absence of the CS and they then established a CS–US association based upon a comparison of these quantities. The comparator hypothesis is a qualitative response rule, which, in principle, can complement any model of acquisition.
Article
Prior research has generally shown that the greater the degree of original learning of a list, the greater the amount of retroactive interference that list suffers. In addition, greater learning of interpolated lists produces more retroactive interference. However, in prior research, the degree of learning has typically been confounded with the amount of retrieval practice on the list. Two free-recall experiments are reported in which subjects studied one original list and then 0, 1, 2, 3, or 4 interpolated lists. The degree of original and of interpolated learning was manipulated by varying exposure time. In Experiment 1, where the typical confounding of retrieval practice and degree of interpolated learning was present, greater interpolated learning induced greater retroactive interference, which is consistent with prior research. However, in Experiment 2, where the degree of interpolated learning was manipulated without concomitant variation in retrieval practice, retroactive interference was the same, whether the interpolated lists had been learned well or poorly. Therefore, greater interpolated learning does not increase the amount of retroactive interference. The results also show that the amount of retroactive interference does not depend on the degree of original learning, in agreement with other work on normal forgetting.
Article
Critical issues in the theoretical and experimental analysis of interference processes in retention are reviewed. The evolution of classical two-factor theory is traced, and the strengths and weaknesses of the contemporary version of this p6sition are examined. Recent critiques of Current interference theories by Martin (1971a) and Greeno, James, and Da Polito (1971) are reviewed and examind. New conceptualizations of interference proposed by these authors, which place major emphasis on retrieval dependencies and on the role of encoding and retrieval processes, are considered and evaluated.
Article
In the first experiment subjects were presented with a number of sets of trials on each of which they could perform a particular action and observe the occurrence of an outcome in the context of a video game. The contingency between the action and outcome was varied across the different sets of trials. When required to judge the effectiveness of the action in controlling the outcome during a set of trials, subjects assigned positive ratings for a positive contingency and negative ratings for a negative contingency. Furthermore, the magnitude of the ratings was related systematically to the strength of the actual contingency. With a fixed probability of an outcome given the action, judgements of positive contingencies decreased as the likelihood that the outcome would occur without the action was raised. Correspondingly, the absolute value of ratings of negative contingencies was increased both by an increment in the probability of the outcome in the absence of the action and by a decrement in the probability of the outcome following the action. A systematic bias was observed, however, in that positive judgements were given under a non-contingent relationship when the outcome frequency was relatively high. However, this bias could be reduced by giving extended exposure to the non-contingent schedule (Experiment 2). This pattern of contingency judgements can be explained if it is assumed that a process of selective attribution operates, whereby people are less likely to attribute an outcome to some potential target cause if another effective cause is present. Experiments 2 and 3 demonstrated the operation of this process by showing that initially establishing another agent as an effective cause of the outcome subsequently reduced or blocked the extent to which the subjects attributed the outcome to the action. Finally, we argue that the pattern and bias in contingency judgements based upon interactions with a causal process can be explained in terms of contemporary conditioning models of associative learning.
Article
In both Pavlovian conditioning and human causal judg- ment, competition between cues is well known to occur when multiple cues are presented in compound and followed by an outcome. More questionable is the occurrence of competition between outcomes when a single cue is followed by multiple outcomes presented in compound. In the experiment reported here, we demonstrated blocking (a type of stimulus competition) between outcomes. When the cue predicted one outcome, its ability to predict a second outcome that was presented in compound with the first outcome was reduced. The procedure mini- mized the likelihood that the observed competition between outcomes arose from selective attention. The competition between outcomes that we observed is problematic for contemporary theories of learning.