ArticlePDF Available

Backward Blocking and Interference Between Cues are Empirically Equivalent in Non-Causally Framed Learning Tasks

Authors:

Abstract and Figures

Backward blocking (BB) and interference between cues (IbC) are cue competition effects produced by very similar manipulations. In a standard BB design both effects might occur simultaneously, which implies a potential problem to study BB. In the present study with humans, the magnitude of both effects was compared using a non causal scenario and a within subjects design. Previous studies have made this comparison using learning tasks framed within causal scenarios. This posits a limit to generalizing their findings to non-causal learning situations because there is ample evidence showing that participants engage in causal reasoning when tasks are causally framed. The results obtained showed BB and IbC effects of the same magnitude in a non causal framed task. This highlights the methodological need for an IbC control in BB experiments.
Content may be subject to copyright.
Backward Blocking 1
Running Head: BACKWARD BLOCKING AND INTERFERENCE BETWEEN
CUES
Backward Blocking and Interference between Cues are empirically equivalent in
non-causally framed learning tasks
David Luque, Joaquín Morís, Cristina Orgaz*, Pedro L. Cobos and Helena Matute*
University of Málaga
*University of Deusto
Address for correspondence:
Pedro L. Cobos
Departamento de Psicología Básica
University of Málaga
Campus de Teatinos
Málaga 29071 Spain
Phone: (+34) 952131087
Fax: (+34) 952132631
p_cobos@uma.es
Backward Blocking 2
Abstract
Backward blocking (BB) and interference between cues (IbC) are cue competition
effects produced by very similar manipulations. In a standard BB design both effects
might occur simultaneously, which implies a potential problem to study BB. In the
present study with humans, the magnitude of both effects was compared using a non
causal scenario and a within subjects design. Previous studies have made this
comparison using learning tasks framed within causal scenarios. This posits a limit to
generalizing their findings to non-causal learning situations because there is ample
evidence showing that participants engage in causal reasoning when tasks are causally
framed. The results obtained showed BB and IbC effects of the same magnitude in a
non causal framed task. This highlights the methodological need for an IbC control in
BB experiments.
Backward Blocking 3
Acknowledgements
Support for this research was provided by Grant SEJ2007-63691/PSIC from
Dirección General de Investigación of the Spanish Government, Grants SEJ-406 and
SEJ-03586 from Junta de Andalucía, and Grant PI2008-9 from the Basque Government.
David Luque and Joaquín Morís were supported by F.P.D.I. fellowships from Junta de
Andalucía. We would like to thank Francisco J. López and Miguel A. Vadillo for their
insightful comments on this experiment. Correspondence concerning this article should
be addressed to Pedro L. Cobos, Departamento de Psicología Básica, Campus de los
Teatinos, Universidad de Málaga, 29071, Málaga, Spain, or to Helena Matute,
Departamento de Psicología, Universidad de Deusto, Apartado 1, 48080 Bilbao, Spain.
E-mail: p_cobos@uma.es or matute@deusto.es.
Backward Blocking 4
Backward Blocking and Interference between Cues are empirically equivalent in
non-causally framed learning tasks
Backward Blocking (BB hereafter) and Interference between Cues (IbC hereafter)
are two learning and memory phenomena with many similarities regarding the standard
experimental conditions in which they are observed. In a BB design, a compound of two
cues, A and B, is first paired with an outcome. Later on, one of the elements of the
compound, B, is presented repeatedly with the same outcome (i.e., ABO1 followed
by BO1). When the other element of the compound, A, is subsequently tested, it
elicits a lower response than if participants had not been exposed to BO1 pairings
(e.g., Shanks, 1985). Similarly, in an IbC design, a cue, A, is paired with an outcome,
and, later on, another cue, B, is trained with the same outcome (i.e., AO1followed by
BO1). As in the case of backward blocking, when Cue A is then presented at test,
participants‟ responses are weaker than in a control group not exposed to B O1
pairings (e.g., Matute & Pineño, 1998). Thus, in both phenomena the BO1
relationship learned during the second stage reduces the expression of the previously
learned relationship between the absent cue, A, and the outcome. Thus, at the empirical
level, the only difference between them is whether the two cues receive compound
training during Stage 1.
This apparently slight difference is a crucial one because, according to the
theoretical accounts of BB, training Cues A and B in compound is a necessary condition
to observe a BB effect (e.g., Aitken, Larkin, & Dickinson, 2001; Dickinson & Burke,
1996; Stout & Miller, 2007). Thus, if the decrease observed in responses to A were
found without presenting A and B together during the first training phase, as in an IbC
Backward Blocking 5
experiment, such effect could not be taken as a case of BB. As a consequence, to infer a
BB effect it should be empirically shown that the compound training of a BB design
adds something to the mere separate training of an IbC design. In other words, as a
methodological requirement, the experimental evidence for BB should be based on the
comparison between BB and IbC in the same experiment. In case of a greater impact of
the BB preparation, a true BB effect can be inferred.
However, because IbC is a relatively recent phenomenon in human contingency
learning, and has only been studied in very few laboratories, researchers working on BB
have rarely used an IbC control condition in BB studies. Thus, BB has been inferred in
many BB studies despite not having provided any evidence supporting the causal role of
compound training of cues. This leads us to the conclusion that many, or at least, some
of the so-called BB effects found so far could actually be cases of IbC.
To our knowledge, there are only two studies comparing BB and IbC in the same
experiment. In one of them, Vadillo, Castro, Matute, and Wasserman (2008) observed a
significantly larger effect of BB than of IbC. The authors interpreted this difference as
evidence of a true BB effect which would be different from an IbC effect. Similarly,
Escobar, Pineño, and Matute (2002) found significant effects of BB and IbC. However,
they reported no difference in size between BB and IbC, which was interpreted as
evidence of similar processes underlying both.
The fact that the instructions used in both experiments suggested a causal
interpretation of the relationships between cues and outcomes makes it difficult to
extrapolate the results to non-causal learning situations. There is considerable evidence
showing that learners engage in causal reasoning processes when the learning tasks are
Backward Blocking 6
framed within causal scenarios (e.g., Waldmann & Holyoak, 1992). Also, it has been
shown that in such circumstances both BB and IbC are modulated by the causal
direction of the learning task (Booth & Buehner, 2007; Cobos, López, & Luque, 2007).
Specifically, BB is observed when learning occurs in the predictive causal direction
(i.e., when the event presented first in each learning trial the cue- is described as the
cause which produces the event presented subsequently the outcome). However, BB
reduces or disappears when learning occurs in the diagnostic causal direction (i.e., when
the cue and the outcome are interpreted as the effect and the cause, respectively).
Conversely, IbC is observed in the diagnostic but not in the predictive causal direction.
Thus, an important implication of those results is that the inconsistency between
Vadillo et al.‟s (2008) and Escobar et al.‟s (2002) results can be explained by a causal
reasoning account of BB and IbB. Specifically, participants in Vadillo et al.‟s (2008)
experiment learned in the predictive direction because the learning task was framed
within a predictive causal scenario in which the cues were described as causes and the
outcomes as effects. In this scenario, participants were instructed to imagine that they
were allergists that had to investigate which foods caused certain allergic reactions in
one of their patients. In each trial, the foods (causes) were presented first and were
followed by the allergic reactions (i.e., effects presented as outcomes). Therefore, the
participants learned in the predictive causal direction. Thus, according to the causal
reasoning account, BB but not IbC should be observed. In fact, Vadillo et al. (2008)
found a greater effect of BB than of IbC, the latter being rather small. This difference
was not observed in Escobar et al.‟s (2002) experiment in which the instructions
suggested a diagnostic causal interpretation of the learning task. In this case, the
participants were instructed to imagine that they had to help a group of refugees to
Backward Blocking 7
escape from a war zone in several trucks. In this task, the cues were light indicators that
signalled the presence or absence of several obstacles (e.g., hidden mines) on the road
which played the role of outcomes (see also Pineño, Ortega, & Matute, 2000). Thus,
participants had to infer whether there were obstacles, as hidden mines, on the road,
from the illumination of indicator lights to decide whether to place or not refugees in the
trucks. Though the instructions did not explicitly state the causal status of the events,
the diagnostic interpretation seems very likely given that people are very familiar with
the existence of mine detectors and devices that can detect metals or other sort of
materials. In any case, it is very unlikely for participants to have attributed any causal
power to the illumination of the lights to produce the different obstacles on the road (see
Cobos et al., 2007, and Luque, Cobos, & López, 2008, for further details on the causal
interpretation of these scenarios).
The considerations above raise an interesting question, namely what would be the
case if BB and IbC were tested in a neutral situation in which none of them is promoted
over the other by suggesting either a predictive or a diagnostic causal interpretation of
the learning task. Testing this idea was therefore the aim of the present study. Thus, in
the present experiment we compared BB and IbC in a situation in which the instructions
suggested arbitrary, non-causal, relationships between cues and outcomes. It is
important to note that, in the context of human causal learning, a positive contingency
between a cue and an outcome does not imply a causal relationship between both
events. Indeed, it has been shown that human beings are capable of differentiating
between mere contingent relationships and causal relationships (see, e.g., Cheng, 1997)
as well as between predictive and causal relationships (e.g., Vadillo & Matute, 2007).
Therefore, in the present research we used colored rectangles as cues and plants as
Backward Blocking 8
outcomes in order to make sure that the relationship between them was arbitrary. The
words cues and outcomes are therefore used only here just to denote the event
presented first (cues: colored rectangles) and the event presented second (outcomes:
plants), without any reference to any potential causal relationship between them.
Moreover, although our task included cues and outcomes with a positive contingency,
the instructions presented to the participants did not use any of these words (causes,
effects, cues, outcomes) and did not suggested any causal role for the events. Thus, we
used a simpler situation to help draw clearer conclusions about the magnitudes of BB
and IbC effects without the potential confound of causal interpretations.
(Table 1 about here)
Method
Participants and apparatus
Forty-two psychology students from the University of Deusto volunteered for this
study. The experiment was run in a computer room with capacity for sixty participants
keeping one-meter distance from each other. The task was performed on personal
computers equipped with home-built software written in Visual Basic 2005 (Microsoft,
USA) with the participants responding via the keyboard. The horizontal distance
between the participants head and the monitor was, approximately, 120 cm.
Procedure
Backward Blocking 9
The design of the experiment is shown in Table 1. During Phase 1, the
participants saw three compound cues (AB, CD and EF) that were paired with three
outcomes (O1, O2 and O3, respectively). Each type of trial in Phase 1 was presented 20
times in pseudorandom order (i.e., avoiding more than two consecutive trials of the
same type). During Phase 2, the participants were exposed to 15 B-O1 trials
pseudorandomly intermixed with 15 G-O3 trials. BB of Cue A should be observed
because, during the first phase, A was always compounded with a competing cue, B,
which, during the second phase, was followed by the same outcome (O1). G was a new
cue that was presented only during Phase 2 and which shared the outcome with the
compound cue EF, as in an IbC design. Therefore, IbC should be observed when testing
either E or F because none of these cues had been trained in compound with G. The
compound cues C and D, and the corresponding outcome, O2, were not presented in
Phase 2 and therefore both C and D could serve as overshadowing controls. In the test
phase, Cues A, C, and E were presented only once per participant in counterbalanced
order. It is important to note that the manipulation was within-subject, hence, all
participants performed the three experimental conditions: BB, IbC as well as the
overshadowing control condition (see Table 1).
The procedure of the experiment was similar to that in Luque, Morís, Cobos and
López (2009) study. First, participants read the instructions and had the opportunity to
ask questions. The participants could earn points by betting on each trial. In order to do
so, they had to learn the relationships between some coloured rectangles and some
fictitious plants which played the role of cues and outcomes, respectively (an English
translation of the task instructions are given in the Appendix section). The coloured
rectangles were blue, brown, yellow, orange, red, green, and pink, and their role as the
Backward Blocking 10
abstract cues shown in Table 1 were counterbalanced. The plants were pictures of
fictitious plants labelled as Kollin, Dobe, and Yamma, and their role as the abstract
outcomes shown in Table 1 were also counterbalanced. The response options were
randomly placed at the left, middle, and right bottom of the screen on a trial-by-trial
basis. For this reason, the first thing to occur immediately before each trial was the
display of the three possible random response options, one per outcome, consisting of
the labelled plant photos. This was done so that participants could know, before the
current trial started, which key should be pressed during that trial for each of the three
possible outcomes. Under each response option, a scroll bar, together with a text box,
was displayed to indicate the amount of points the participant was betting for the
corresponding option in each trial. Then, the cue, or cues, (i.e., one or two of the
coloured rectangles) appeared at the middle top of the screen for 2.5 s during which the
participants had to bet which of the three plants they thought was related with the cue
(or cues) that were present in the screen in that trial. When two cues were displayed on
the same trial, they were placed one beside the other, the specific location for each one
being counterbalanced on a trial-by-trial basis. Once the 2.5 s time had passed, the cue
(or cues) disappeared, which was indicated by the rectangle (or rectangles) taking on the
grey colour (see Figure 1). To respond, the participants placed their bets by pressing
either Key “1”, or Key “2”, or Key “3” for the plant (i.e., the response option) placed at
the left, middle, or right bottom of the screen, respectively. While a given response key
was kept pressed, the points bet for the corresponding option increased continuously,
which was indicated analogically, by the movement of the scroll-bar from left to right,
as well as with numbers in the corresponding text box ranging from 0 to 100. Once the
cue had disappeared, pressing any of the response keys had no effect on the amount of
Backward Blocking 11
points bet. On each trial, participants earned as many points as those bet for the correct
outcome, and lost as many points as those bet for an incorrect outcome. After each bet,
feedback was given consisting of: a) the correct plant, which was indicated by keeping
it visible, and removing the remaining ones from the screen, b) the amount of points
earned on a given trial, which was indicated in a text box placed on the centre of the
screen over the labelled plant photos, and c) the total points gained throughout the
training trials, which was indicated in a text box at the right top of the screen.
The test phase consisted of additional trials which only differed from the previous
training trials in two respects: cues were presented for 5s, and the participants received
no feedback after their bets.
(Figure 1 about here)
Analysis
To make sure that the data corresponded to participants who had understood the
instructions and paid attention to the task, their total scores were first descriptively
analyzed. The data from those participants whose total score at the end of the two
learning phases were two standard deviations below or above the mean of the whole
sample were removed. Using this criterion two participants that scored two standard
deviations below the mean were excluded from the analysis. In addition, due to a
software error, the data from six participants were also removed from the analysis. The
test phase for these participants included two Cue C trials and no Cue A trials. Thus, we
could not collect data regarding the BB condition from these participants. All the
statistical analyses were performed for the remaining participants (N = 34). The
dependent measure for these analyses was the number of points bet on the correct
Backward Blocking 12
outcome in test trials -e.g., in the test for BB, the dependent measure was the number of
points bet on Outcome O1 (see Table 1). As an additional dependent measure, we
calculated the accuracy of responses in each condition, i.e., the percentage of correct
responses.
Results
As shown in Figure 2, responding to Cue E was weaker than responding to Cue
C, which suggests that IbC was found. BB is also evident in that responding to Cue A
was weaker than responding to Cue C. Moreover, both the BB and the IbC effects
seemed to be of similar size. A repeated measures ANOVA performed on participants‟
responses at test confirmed these impressions, revealing a significant effect of cue (A,
C, E), F(2, 66) = 5.03, MSE = 1609.5, p = .009, η2 = 0.13. Least Significant Difference
(LSD) pairwise comparisons showed a significant IbC effect (cue E vs. C, p = .019, η2 =
0.16) as well as a significant BB effect (cue A vs. C, p = .002, η2 = 0.24). Finally, no
differences were detected between both effects (cue A vs. E, p = .630, η2 < 0.01).
Additionally, we repeated the same analysis with another dependent variable in
which the accuracy of responses was also assessed. To achieve this aim, we analyzed
the percentage of correct responses at test in each experimental condition. The
percentage of correct responses was 53.79%, 52.14% and 81.33% for the conditions
IbC, BB and Overshadowing, respectively. Also, a repeated measures ANOVA was
performed on participants‟ percentage of correct responses at test, revealing a
significant effect of condition (IbC, BB, Overshadowing), F(2, 66) = 4.28, MSE =
2134.5, p = .018, η2 = 0.12. LSD pairwise comparisons showed the same pattern than in
the previous analysis, i.e., a significant IbC effect (IbC vs. Overshadowing, p = .013, η2
Backward Blocking 13
= 0.17), a significant BB effect (BB vs. Overshadowing, p = .005, η2 = 0.21), as well as
no differences between the IbC and the BB conditions (p = .9, η2 < 0.01).
(Figure 2 about here)
Discussion
BB and IbC are very similar effects that are usually accounted for by very
different associative accounts. In both cases, responding to the target cue is reduced by
later pairings of an alternative cue with the same outcome. The only difference between
both effects is that, while BB requires that the target and the alternative cue be
previously trained in compound, IbC is found when both cues are trained apart (see
Table 1). Interestingly, although to infer a true BB effect it is necessary to empirically
show that training cues in compound makes a difference compared to a mere IbC
preparation, this has been rarely taken into account for control purposes in BB
experiments.
To our knowledge, the use of an IbC control in BB experiments has only been
carried out in two previous studies in which BB and IbC were compared in the same
experiment (Escobar et al., 2002; Vadillo et al., 2008). Whereas BB was found to be
significantly greater than IbC in Vadillo et al. (2008), Escobar et al. (2002) reported
observing BB and IbC effects of similar magnitude. However, both experiments were
based on learning tasks framed within causal scenarios that could have made
participants engage in causal reasoning processes. In fact, as explained in the
Backward Blocking 14
Introduction, the difference between Vadillo et al.‟s (2008) and Escobar et al.‟s (2002)
results is consistent with a causal reasoning account of BB and IbC if we assume that
participants learned in the cause-effect direction in the former case, and in the effect-
cause direction in the latter case (Booth & Buehner, 2007; Cobos et al., 2007). Thus, the
results found so far cannot be straightforwardly extrapolated to human contingency
learning situations in which no causal scenario is used, and, thus, no effect is promoted
over the other on the basis of causal reasoning processes.
As in previous studies, the present study used a design intended to compare BB
and IbC in the same experiment. Unlike previous experiments our task instructions did
not suggest any causal interpretation of the learning task. Thus, the use of causal
reasoning processes was prevented because such mechanisms are only evoked in
situations that are clearly interpretable as causal (Lopez, Cobos, & Caño, 2005).
Our results showed a significant effect of both IbC and BB. Moreover, both
effects were of equivalent magnitude. Thus, when no causal interpretation is suggested
through instructions, BB and IbC seem to be empirically undistinguishable, at least
when a standard BB and IbC design is used. It seems therefore sensible to suggest that
in the absence of causal reasoning processes, compound training of cues do not seem to
add anything to separate training of cues. This, in turn shows the difficulties of inferring
BB when the appropriate control is used in situations in which participants do not
engage in causal reasoning processes.
Of course, because no independent test was used to assess whether participants
gave a causal interpretation to cues and outcomes, one could still think that participants
could have engaged in causal reasoning processes. However, this possibility seems quite
Backward Blocking 15
unlikely. Interpreting coloured rectangles as causing the presence of plants or the
presence of plants as making rectangles take on different colours in the context of a
gambling task would require something more than merely labelling cues and outcomes
as cause or effect. Such interpretation has to be framed within the context of a causal
scenario based on plausible causal laws and causal mechanisms. However, given that
the instructions did not suggest any causal scenario at all, a good amount of cognitive
load would have been necessary to come up with causal laws or causal mechanisms
connecting cues and outcomes, not to mention the questionable benefit from doing this
to face the learning task.
Accepting that causal inference processes are very unlikely to have been involved,
we could explain our result in associative terms. The associative explanations of BB are
based on the role played by within-compound associations. According to these
explanations, BB should not be observed in the absence of within-compound
associations (e.g., Dickinson & Burke, 1996). Although the BB effect obtained in our
experiment is a straightforward prediction of these models, the IbC effect observed is
clearly not predicted because the interfering cue presented during the second phase was
never trained in compound during the first phase with the interfered cue. On the other
hand, if we assume that participants encoded the AB, EF, and CD compounds as unitary
configurations (not reducible to its elements, e.g., Pearce, 1994), the BB treatment
would become X-O1, B-O1 (where X stands for the representation of the compound),
and the IbC treatment would become Y-O3, G-O3 (where Y stands for the
representation of the compound). Hence, these two training conditions would be
equivalent IbC conditions and the mechanisms underlying IbC would cause a response
decrement in both conditions.
Backward Blocking 16
It is important to note, however, that before speculating on a common
interpretation of BB and IbC, we need to gather more compelling evidence regarding
the relationship between both phenomena. This would certainly involve the use of
dissociation strategies, which are more suitable to test whether the two phenomena are
produced by the same processes or not. At the moment, our study is one step before in
that it clearly shows the potential confounding between BB and IbC in non-causal
learning tasks, leaving open the question of whether most of the BB results reported so
far are true instances of BB or not. Hence, our study points to the need of controlling for
IbC in BB experiments.
Thus, although there are numerous studies that have reported BB effects (e.g.,
Dickinson & Burke, 1996; Shanks, 1985; Van Hamme & Wasserman, 1994;
Wasserman & Berglan, 1998), very few compared BB against IbC, and among those
that did compare both effects (Escobar et al., 2002; Vadillo et al., 2008), none of them
used a situation which makes causal reasoning processes unlikely. In some cases,
although the experiments conducted have not been intended to explicitly compare BB
with an IbC control, the design used can be easily regarded as well-suited for that
purpose. For example, Larkin, Aitken, & Dickinson (1998) reported three experiments
in which they failed to observe any evidence for BB when they used a control condition
that was actually equivalent to our IbC condition. They used several overshadowing
controls sharing the same outcome, e.g., GH-O1, IJ-O1, which could be considered as
equivalent to our IbC condition (see Table 1). Thus, their results could be readily
interpreted as showing, like the present ones, effects of BB and IbC which are
empirically identical. The only case in which they reported a BB-like effect was when
their control consisted of presentations of Cue B predicting the absence of the outcome
Backward Blocking 17
during Phase 2 (i.e., a condition called recovery from overshadowing because it tends to
enhance, rather than reduce, the response to the target cue).
In summary, the main contribution of the present study is the finding of
empirically equivalent effects of BB and IbC when causal reasoning processes are
unlikely to be involved. Further experiments should be conducted to find out whether
these effects are empirically distinguishable in other experimental conditions.
Backward Blocking 18
References
Aitken, M. R. F., Larkin, M. J. W., & Dickinson, A. (2001). Re-examination of the role
of within-compound associations in the retrospective revaluation of causal
judgements. Quarterly Journal of Experimental Psychology, 54B, 2751.
Booth, S. L., & Buehner, M. J. (2007). Asymmetries in cue competition in forward and
backward blocking designs: Further evidence for Causal Model Theory. Quarterly
Journal of Experimental Psychology, 60(3), 387399.
Cheng, P. W. (1997). From covariation to causation: A causal power theory.
Psychological Review, 104, 367405.
Cobos, P. L., López, F. J., & Luque, D. (2007). Interference between cues of the same
outcome depends on the causal interpretation of the events. Quarterly Journal of
Experimental Psychology, 60(3), 369386.
Dickinson, A., & Burke, J. (1996). Within-compound associations mediate the
retrospective revaluation of causality judgements. Quarterly Journal of
Experimental Psychology, 49B, 6080.
Escobar, M., Pineño, O., & Matute, H. (2002). A comparison between elemental and
compound training of cues in retrospective revaluation. Animal Learning &
Behavior, 30, 228238.
Backward Blocking 19
Larkin, M. J., Aitken, M. R. F., & Dickinson, A. (1998). Retrospective revaluation of
causal judgments under positive and negative contingencies. Journal of
Experimental Psychology: Learning, Memory and Cognition, 24, 13311352.
López, F. J., Cobos, P. L., & Caño, A. (2005). Associative and causal reasoning
accounts of causal induction: symmetries and asymmetries in predictive and
diagnostic inferences. Memory & Cognition, 33, 13881398.
Luque, D., Cobos, P. L., & López, F. J. (2008). Interference between cues requires a
causal scenario: Favourable evidence for causal reasoning models in learning
processes. Learning and Motivation, 39, 196208.
Luque, D., Morís, J., Cobos, P. L., & López, F. J. (2009) Interference between cues of
the same outcome in a non-causally framed scenario. Behavioural Processes, 81,
328332.
Matute, H., & Pineño, O. (1998). Stimulus competition in the absence of compound
conditioning. Animal Learning & Behavior, 26, 314.
Pearce, J. M. (1994). Similarity and discrimination: a selective review and a
connectionist model. Psychological Review, 94, 6173.
Pineño, O., Ortega, N., & Matute (2000). The relative activation of associations
modulates interference between elementally trained cues. Learning and
Motivation, 31, 128152.
Shanks, D. R. (1985). Forward and backward blocking in human contingency
judgement. Quarterly Journal of Experimental Psychology, 37B, 121.
Backward Blocking 20
Stout, S. C., & Miller, R. R. (2007). Sometimes competing retrieval (SOCR): A
formalization of the comparator hypothesis. Psychological Review, 114, 759783.
Vadillo, M. A., Castro, L., Matute, H., & Wasserman, E. A. (2008). Backward
blocking: The role of within-compound associations and interference between
cues trained apart. Quarterly Journal of Experimental Psychology, 61, 185193.
Vadillo, M. A., & Matute, H. (2007). Predictions and causal estimations are not
supported by the same associative structure. Quarterly Journal of Experimental
Psychology, 60, 433-447.
Van Hamme, L. J., & Wasserman, E. A. (1994). Cue competition in causality
judgments: The role of nonpresentation of compound stimulus elements. Learning
and Motivation, 25, 127151.
Waldmann, M. R., & Holyoak, K. J. (1992). Predictive and diagnostic learning within
causal models: asymmetries in cue competition. Journal of Experimental
Psychology: General, 121, 222236.
Wasserman, E. A., & Berglan, L. R. (1998). Backward blocking and recovery from
overshadowing in human causal judgment: The role of within-compound
associations. Quarterly Journal of Experimental Psychology, 51B, 121138.
Backward Blocking 21
Figure Captions
Figure 1. Information displayed on a training trial, as seen during the actual task. (A)
Cues presentation. The cues were presented in the two rectangles at the top of the
screen. The darker grey shaded rectangle represents a cue presented on the current trial,
i.e., this figure represents a single cue trial. Participants could only respond during the
2.5 s. interval in which the rectangle stayed colored (i.e., cue present). The pictures of
plants at the bottom of the screen represent the three possible response options. The
scrollbars were used to bet points on each outcome (see main text). After this period of
time, the outcome screen was automatically presented. (B) Outcome presentation. The
photos of the incorrect plants and their names disappeared jointly with the cue or cues.
Only the photo of the correct plant (the outcome) remained on the screen with its name
and the number of points gained or lost. Participants had to press the „„X” key to
continue with the next trial.
Figure 2. Mean responses to cues A (Backward Blocking), C (Overshadowing control)
and E (Interference between Cues). Error bars represent standard errors of the mean.
Backward Blocking 22
Table 1. Design summary of the Experiment
Phase 1 (20 trials)
Phase 2 (15 trials)
Test Phase (one ea)
ABO1
CDO2
EFO3
BO1
-
GO3
A? C? E?
Backward Blocking 23
Figure # 1
Backward Blocking 24
Figure # 2
Backward Blocking 25
Appendix
Instructions (Translated from Spanish)
In the task that you are going to do next you will have to earn as many points as
possible. To do so, please read the following instructions with great care and attention.
The task is divided in a series of temporal intervals or trials. During these trials you will
have to learn the relationships between some coloured rectangles and the pictures of
three plants. In each trial you must learn which of the three plants is associated with the
coloured rectangles that appear at the top of the screen. The three plants will be visible
since the beginning of each trial. To add points to your score you must press the key
that corresponds to the plant that you believe is related to the colour of the rectangles on
top. If you press the key related with another plant, these points are going to be
subtracted from your total score.
During the task, in each trial, you will find two rectangles at the top of the screen. They
will be grey at the beginning of the trial, and could then become coloured (one or both
of them), during a few seconds. During the seconds they are coloured, you must decide
which plant is related with these colours. If you believe that it is the plant which is on
the left, you must push the “1” key on the keyboard. If you believe that it is the plant
which is on the center, you must push the “2” key on the keyboard. If you believe that it
is the plant which is on the right, you must push the “3” key on the keyboard. The more
you press the keys the more point you could win (or lose) in each trial. Moreover, the
points that you can earn or lose could rise faster if you maintain the keys pressed down.
It is important that you press the key corresponding to the plant related with the
coloured rectangles at the top of the screen. If you make a mistake, the points that you
Backward Blocking 26
bet in that trial will be subtracted from your total score. In each trial you can see the
points you are betting for each plant in a lateral bar, as well as a counter next to each
plant. After pressing (or not pressing) the keys “1”, “2” or “3” you will see the
following information: (1) The number of points that you have won (or lost) in that trial.
(2) The right plant (that is, the plant you should have bet more points on). You will
know that this one is the right plant because it will be the only one visible in the screen
right after you place your bet. The wrong plants will disappear so you can see which is
the correct one. Therefore you will know if you are doing it right.
For your convenience, the total number of accumulated points can always be seen in the
right top corner of the screen. To continue to another trial, press the X key.
It is important to keep in mind that the location of each plant in the screen may change
from trial to trial. The colours are associated with the plant, not with its location. Thus,
you have to memorize the “colour-plant” association in order to earn points, not the
“colour-key” association (1, 2, or 3). The reason for that is that the plants changes their
location, so, if you want to bet points for one plant you should press different keys in
each trial.
In sum, if you want to earn points you should learn the “colour – plant” relationships
and press the key corresponding to the correct position. The “1”, “2” and “3” keys can
be pressed while the rectangles remain in colour (not grey).
IT IS VERY IMPORTANT that the index finger of your right hand remains over the
“1” key, the middle finger of your right hand over the “2” key and the ring finger of
your right hand over the “3” key during the whole experiment. If you do not do it like
this, you will have to look for the keys each time you want to press them.
Backward Blocking 27
At the end of the experiment, in some of the latest trials, you will not have access to the
information about which plant was the correct one after you place your bet. After a test
trial like this one, just continue performing the task normally.
REMEMBER: Your mission is to learn the relationship between the rectangles of
different colours and the plants in order to earn the maximum possible number of
points.
If you have any doubt, please, ask us!
... Backward [cue] blocking (e.g., Shanks, 1985) and two-phase retroactive cue interference (e.g., Luque et al., 2009Luque et al., , 2010Luque et al., , 2011Matute & Pineño, 1998b;Miguez et al., 2012) together mirror the parallel between [forward] blocking and associative proactive cue interference, but the compound training (i.e., YX-O) of blocking and elemental target training (X-O) of cue interference occurs in Phase 1 of treatment rather than during Phase 2, and the Y-O pairings occur during Phase 2 rather than Phase 1. Specifically, in a backward blocking design, a compound of Y and X is initially paired with an outcome (O) during the first phase of training, and subsequently, elemental pairings of Y and O occur (i.e., XY-O pairings in Phase 1 and Y-O pairings in Phase 2). When X is subsequently tested, it elicits weaker conditioning responding than does X in a control condition in which no Y-O pairings are administered in Phase 2. (Note that backward blocking is a form of retrospective revaluation and is generally a less robust phenomenon than is forward blocking.) ...
... Some prior research has assessed the contribution of retroactive interference between cues to backward blocking in human subjects (e.g., Escobar et al., 2002;Luque et al., 2011;Vadillo et al., 2008; see General Discussion for a brief review of these articles). However, to our knowledge, a systematic assessment of the contribution of proactive associative interference to forward blocking has never been conducted. ...
... However, direct comparison of the two studies is not central to the present conclusions. Luque et al. (2011) found backward blocking and retroactive cue interference to be of similar magnitude and, based on that, suggested that the two phenomena arose from the same underlying process. The present data-specifically, the context dependency of backward blocking-suggest that much of the backward blocking effect is attributable to retroactive cue interference. ...
Article
Full-text available
Blocking (i.e., reduced responding to cue X following YX-outcome pairings in Phase 2 as a consequence of cue Y having been paired with the outcome in Phase 1) is one of the signature phenomena in Pavlovian conditioning. Its discovery promoted the development of multiple associative models, most of which viewed blocking as an instance of pure cue competition (i.e., a decrease in responding attributable to training two conditioned stimuli in compound). Two experiments are reported in which rats were examined in a fear conditioning paradigm (i.e., lick suppression), and context dependency of retrieval at test was used as an index of associative cue interference (i.e., a decrease in responding to a target cue as a result of training a second cue with the same outcome but without concurrent presentation of the two cues). Specifically, we observed renewal of forward-blocking which parallels renewal of proactive interference, and renewal of backward-blocking which parallels renewal of retroactive interference. Thus, both backward-blocking (Experiment 1, embedded in a sensory preconditioning design) and forward-blocking (Experiment 2, conducted in first-order conditioning) appear to be influenced by retroactive and proactive interference, respectively, as well as cue competition. Consequently, blocking, long regarded as a benchmark example of pure cue competition, is sometimes a hybrid of cue competition and associative interference. Finally, the Discussion considers whether stimulus competition and associative interference are two independent phenomena or products of a single underlying process. (PsycInfo Database Record (c) 2022 APA, all rights reserved).
... We expected to observe more interference in groups that received higher numbers of A-O interference trials in Phase 2 (e.g., Group Int 12). Having a larger number of reinforced trials in Phase 2 is analogous to providing extra training (i.e., associative inflation) to the non-target cue after target training, which in cue competition can produce higher competition (i.e., a decrease in responding to the target cue; Balleine, Espinet, & Gonzalez, 2005;Denniston et al., 1996;Luque, Morís, Orgaz, Cobos, Matute;Vadillo, Castro, Matute, & Wasserman, 2008). ...
... Presumably, the associative status of the interfering association was greater with more A-O trials. The observed decrease in behavioral control by X when the interfering cue A was further reinforced parallels retrospective revaluation phenomena often reported in cue competition when there is additional reinforcement of the competing cue by itself (e.g., Balleine et al., 2005;Denniston et al., 1996;Luque et al., 2011;Vadillo et al., 2008). ...
Article
Full-text available
Two fear-conditioning experiments with rats assessed whether retrospective revaluation, which has been observed in cue competition (i.e., when compounded cues are followed with an outcome), can also be observed in retroactive cue interference (i.e., when different cues are reinforced in separate phases with the same outcome). Experiment 1 found that after inducing retroactive cue interference (i.e., X-outcome followed by A-outcome), nonreinforced presentations of the interfering cue (A) decreases interference with responding to the target cue (X), just as has been observed in retrospective revaluation experiments in cue competition. Using the opposite manipulation (i.e., adding reinforced presentations of A), Experiment 2 demonstrated that after inducing retroactive cue interference, additional reinforced presentations of the interfering cue (A) increases interference with responding to the target cue (X); alternatively stated, the amount of interference increases with the amount of training with the interfering cue. Thus, both types of retrospective revaluation occur in retroactive cue competition. The results are discussed in terms of the possibility that similar associative mechanisms underlie cue competition and cue interference.
... To overcome this potential confound, similar analyses were conducted with two alternative measures of response at test: (a) the difference between the number of correct and incorrect responses at test and (b) the percent of correct responses at test. Both variables were used in previous research on human contingency learning (Rosas, Vila, Lugo, & López, 2001; Luque, Morís, Orgaz, Cobos, & Matute, 2011). The means and standard errors of these measures are inTable 2 . ...
... For instance, the present task has been effective in illustrating other cue selection phenomena such as interference between cues (Luque, et al., 2009), which are known to be somewhat elusive (Lipp & Dal Santo, 2002; Luque, Cobos, & López, 2008). The same task has also been used to obtain backward blocking in the past (Luque, Morís, & Cobos, 2010; Luque, et al., 2011). As interference between cues and backward blocking are cue selection phenomena difficult to obtain on some tasks, it seems that our noncausal task is a sensitive in measuring cue selection. ...
Article
Full-text available
Two types of theories are usually invoked to account for cue-interaction effects in human-contingency learning, performance-based theories, such as the comparator hypothesis and statistical models, and learning-based theories, such as associative models. Interestingly, the former models predict two important cue-interaction effects, forward and backward blocking, should affect responding in a similar manner, whereas learning-based models predict the effect of forward blocking should be larger than the effect of backward blocking. Previous experiments involved important methodological problems, and results have been contradictory. The present experiment was designed to explore potential asymmetries between forward and backward blocking. Analyses yielded similar effect sizes, thereby favoring the explanation by performance-based models.
... The inference process proposed in our structural belief hypothesis is cognitively demanding, and might contribute to the RIBC effect, particularly when experimental conditions favor the deployment of multiple cognitive resources. For instance, the operation of highly demanding reasoning processes should be particularly favoured in experiments in which there are only a few deterministic relationships to be learned and without time pressure to respond during the test (e.g.,Escobar et al., 2002;Luque et al., 2008Luque et al., , 2010Luque et al., , 2011Matute & Pineno, 1998a). It should be noted that this is not the case in the experiments published in the paired-associate learning literature (e.g,Keppel et al., 1971). ...
Article
The effect of retroactive interference between cues predicting the same outcome (RIBC) occurs when the behavioral expression of a cue–outcome association (e.g., A→O1) is reduced due to the later acquisition of an association between a different cue and the same outcome (e.g., B→O1). In the present experimental series, we show that this effect can be modulated by knowledge concerning the structure of these cue–outcome relationships. In Experiments 1A and 1B, a pretraining phase was included to promote the expectation of either a one-to-one (OtO) or a many-to-one (MtO) cue–outcome structure during the subsequent RIBC training phases. We hypothesized that the adoption of an OtO expectation would make participants infer that the previously learned A→O1 relationship would not hold any longer after the exposure to B→O1 trials. Alternatively, the adoption of an MtO expectation would prevent participants from making such an inference. Experiment 1B included an additional condition without pretraining, to assess whether the OtO structure was expected by default. Experiment 2 included control conditions to assess the RIBC effect and induced the expectation of an OtO or MtO structure without the addition of a pretraining phase. Overall, the results suggest that participants effectively induced structural expectations regarding the cue–outcome contingencies. In turn, these expectations may have potentiated (OtO expectation) or alleviated (MtO expectation) the RIBC effect, depending on how well these expectations could accommodate the target A→O1 test association. This pattern of results poses difficulties for current explanations of the RIBC effect, since these explanations do not consider the incidence of cue–outcome structural expectations.
... This is possibly due to the fact that the filler cues presented during Phase 3 are paired with the same outcomes used for the target cue x. It is well known that training a novel cue-outcome association hinders responding to other cues that have been paired with the same outcome in the past (i.e., interference between cues; see Escobar et al., 2002;Vadillo et al., 2008;Luque et al., 2009Luque et al., , 2011. This interference effect might have limited overall responding to x at test, which would explain why some of the key statistical contrasts failed to reach conventional levels of significance. ...
Article
Full-text available
Decades of research in extinction and interference show that contexts can play a critical role at disambiguating the meaning of cues that have been paired with different outcomes at different times. For instance, if a cue x is followed by outcome 1 in the first phase of an experiment and by outcome 2 in a second phase, responses to cue x tend to be consistent with outcome 2 when tested in a context similar to that of the second phase of the experiment. However, if participants are taken back to the original context of the first phase (i.e., ABA renewal) or to a completely new context (i.e., ABC or AAB renewal), their responses to x tend to be more consistent with outcome 1. Although the role of physical and temporal contexts has been well studied, other factors that can also modulate the selective retrieval of information after interference have received less attention. The present series of experiments shows how changes in cue configuration can modulate responding in a similar manner. Across five experiments using a human predictive learning task, we found that adding, removing or replacing elements from a compound cue that had undergone an interference treatment gave rise to a recovery of responding akin to that observed after context changes in AAB renewal. These results are consistent with those of previous studies exploring the effect of changes of cue configuration on interference. Taken together, these studies suggest that a change in cue configuration can have the functional properties of a context change, a finding with important implications for formal models of configural learning and for classical accounts of interference and information retrieval.
... These effects have been extensively studied, and several theoretical interpretations have been advanced (for a review, see Bouton, 2010). On the other hand, however, some experiments have explored A-B, C-B interference effects ( Matute and PineñoPine˜Pineño, 1998; see also Luque et al., 2011;Vadillo et al., 2008a,b). ...
Article
a b s t r a c t Some researchers have attempted to determine whether situations in which a single cue is paired with several outcomes (A–B, A–C interference or interference between outcomes) involve the same learning and retrieval mechanisms as situations in which several cues are paired with a single outcome (A–B, C–B interference or interference between cues). Interestingly, current research on a related effect, which is known as retrieval-induced forgetting, can illuminate this debate. Most retrieval-induced forgetting experiments are based on an experimental design that closely resembles the A–B, A–C interference paradigm. In the present experiment, we found that a similar effect may be observed when items are rearranged such that the general structure of the task more closely resembles the A–B, C–B interference paradigm. This result suggests that, as claimed by other researchers in the area of contingency learning, the two types of interference, namely A–B, A–C and A–B, C–B interference, may share some basic mecha-nisms. Moreover, the type of inhibitory processes assumed to underlie retrieval-induced forgetting may also play a role in these phenomena.
Article
Full-text available
Most associative theories have assumed that stimulus competition occurs only between conditioned stimuli (CSs) that are trained in compound. The present research investigated the possibility of competition between two CSs that were individually paired to the same unconditioned stimulus (US). We used human subjects in an anticipatory suppression analogue to Pavlovian conditioning. Experiment 1 showed that X+ training followed by A+ training resulted in impaired responding to X. This did not occur when A+ training preceded X+ training. Experiment 2 replicated the basic effect and showed that it did not occur when the Phase 2 training consisted of A− instead of A+ nor when the A+ pairings occurred in a second context. Experiment 3 showed that A+ pairings occurring in a second context could still produce the effect when X was tested in the context in which the A+ pairings had occurred, but not when X was tested in a context different from that used for A+ training. Collectively, these results suggest that individually trained CSs may compete with each other when one of those CSs is more strongly activated by the test context than the other one.
Article
Full-text available
Because causal relations are neither observable nor deducible, they must be induced from observable events. The 2 dominant approaches to the psychology of causal induction-the covariation approach and the causal power approach-are each crippled by fundamental problems. This article proposes an integration of these approaches that overcomes these problems. The proposal is that reasoners innately treat the relation between covariation (a function defined in terms of observable events) and causal power(an unobservable entity) as that between scientists' law or model and their theory explaining the model. This solution is formalized in the power PC theory, a causal power theory of the probabilistic contrast model(P. W. Cheng & L. R. Novick, 1990). The article reviews diverse old and new empirical tests discriminating this theory from previous models, none of which is justified by a theory. The results uniquely support the power PC theory.
Article
Retrospective revaluation of causal judgments was investigated in a 2-stage procedure. In the 1st stage, compounds of 2 cues were associated with the outcome, whereas in the 2nd stage, a cue from each compound was trained by itself. Associating this cue with the outcome in the 2nd stage had no detectable effect on the causal rating of the other cue from the compound, whereas presenting it without the outcome enhanced the causal rating of the other cue. The retrospective revaluation of the causal rating of these productive cues and also of preventative cues depended on consistent pairing of the cues during compound training, suggesting a role for within-compound associations. These results favor associative accounts of retrospective revaluation that use separate excitatory and inhibitory learning processes rather than a general error-correcting learning algorithm.
Article
College students rated the causal efficacy of Elements X, A, and B of food compounds AX and BX in producing the allergic reaction of a hypothetical patient. The results of a 16-day allergy test were presented to subjects in a serial, trial-by-trial manner. The response format used was a running estimate, in which subjects were asked to rate all of the three foods after each of the 16 trials. Ratings of distinctive Elements A and B diverged and ratings of common Element X decreased as the difference in the correlation of AX and BX with the occurrence and nonoccurrence of the allergic reaction increased. These human causal judgments closely correspond with stimulus selection effects observed in the conditioned responses of animals in associative learning studies. The experiment also directly demonstrated the fact that significant changes in the causal ratings of a stimulus occur on trials in which the cue is not presented. Associative theories such as that of Rescorla and Wagner (1972) predict changes in associative strength only for those stimulus elements that are presented on a particular trial. A modification of the Rescorla-Wagner model is described that correctly predicts immediate changes in the associative strengths of all relevant cues on each trial—whether presented or not.
Article
Three experiments investigated whether a process akin to L. J. Kamin's (1969) blocking effect would occur with human contingency judgments in the context of a video game. 102 students were presented with sets of trials on each of which they could perform a particular action and observe whether the action produced a particular outcome in a situation in which there was an alternative potential cause of the outcome. Exp I showed that prior observation of the relationship between the alternative cause and the outcome did indeed block or reduce learning about the subsequent action-outcome relationship. However, exposure to the relationship between the alternative cause and the outcome after observing the association between the action and the outcome also reduced judgments of the action-outcome contingency (backward blocking), a finding at variance with conditioning theory. In Exp II, it was found that the degree of backward blocking depended on the predictive value of the alternative cause. Finally, Exp III showed that the backward blocking effect was not the result of greater forgetting about the action-outcome relationship in the experimental than in the control condition. Results cast doubt upon the applicability of contemporary theories of conditioning to human contingency judgment.
Article
Retrospective revaluation of causal judgments was investigated in a 2-stage procedure. In the 1st stage, compounds of 2 cues were associated with the outcome, whereas in the 2nd stage, a cue from each compound was trained by itself. Associating this cue with the outcome in the 2nd stage had no detectable effect on the causal rating of the other cue from the compound, whereas presenting it without the outcome enhanced the causal rating of the other cue. The retrospective revaluation of the causal rating of these productive cues and also of preventative cues depended on consistent pairing of the cues during compound training, suggesting a role for within-compound associations. These results favor associative accounts of retrospective revaluation that use separate excitatory and inhibitory learning processes rather than a general error-correcting learning algorithm. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
In an interference-between-cues design (IbC), 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. The present study examined whether IbC could be caused by associative mechanisms independent of causal reasoning processes. This was achieved by testing participants in two different learning situations. In the Causal Scenario condition, participants learned in a diagnostic situation in which a common cause (Outcome 1) caused two disjoint effects, namely Cues A and B. In the Non-Causal Scenario condition, the same IbC design and stimulus conditions were used. However, instructions provided no causal frame to make sense of how cues and outcomes were related. IbC was only found in the Causal Scenario condition. This result is consistent with Causal Reasoning Models of causal learning and raises important difficulties for associative explanations of IbC.
Article
Matute and Pineño (1998a) showed evidence of interference between elementally trained cues and suggested that this effect occurs when the interfering association is more strongly activated than the target association at the time of testing. The present experiments tested directly the role of the relative activation of the associations in the effect of interference between elementally trained cues. In three human experiments we manipulated the relative activation of the interfering and target associations in three different ways: (a) introducing a retention interval between training of the interfering association and the test trial (Experiment 1); (b) training the target and the interfering associations in a single phase, instead of training them in separate phases (Experiment 2); and (c) introducing, just before testing, a novel cue which, like the retention interval used in Experiment 1, had the purpose of separating the interfering trials from the test trial (Experiment 3). All three manipulations led to an enhancement of responding to the target association at testing, suggesting that they were effective in preventing the interfering association from being the most strongly activated one at the time of testing. Taken together, these results add further evidence on how the relative activation of associations modulates interference between elementally trained cues.
Article
Retroactive interference between cues of the same outcome (i.e., IbC) occurs when the behavioral expression of an association between a cue and an outcome (e.g., A-->O1) is reduced due to the later acquisition of an association between a different cue and the same outcome (e.g., B-->O1). Though this interference effect has been traditionally explained within an associative framework, there is recent evidence showing that IbC effect may be better understood in terms of the operation of higher order causal reasoning processes. The results from Experiments 1 and 2 showed an IbC effect in a learning task within a game scenario suggesting non-causal relationships between events. Thus, these results showed that IbC may have a diverse origin, one of them being of an associative nature.
Article
Several researchers have recently claimed that higher order types of learning, such as categorization and causal induction, can be reduced to lower order associative learning. These claims are based in part on reports of cue competition in higher order learning, apparently analogous to blocking in classical conditioning. Three experiments are reported in which subjects had to learn to respond on the basis of cues that were defined either as possible causes of a common effect (predictive learning) or as possible effects of a common cause (diagnostic learning). The results indicate that diagnostic and predictive reasoning, far from being identical as predicted by associationistic models, are not even symmetrical. Although cue competition occurs among multiple possible causes during predictive learning, multiple possible effects need not compete during diagnostic learning. The results favor a causal-model theory.