|The Other Emergent Performance: Recent Basic, Translational, and Applied Research Advances in Exclusion Learning|
|Tuesday, May 31, 2016|
|2:00 PM–3:50 PM |
|Zurich FG, Swissotel|
|Area: EAB/PRA; Domain: Translational|
|Chair: Richard W. Serna (University of Massachusetts Lowell)|
|Discussant: William J. McIlvane (University of Massachusetts Medical School)|
Teaching methods based on exclusion learning are effective for rapidly establishing conditional discriminations under a variety of applications. Beyond its effectiveness, the exclusion-learning procedure is theoretically interesting, as it not only results in emergent performance, but the stimulus-control requirements of the task offer opportunities to examine the fundamental nature of conditional-discrimination learning. Moreover, the exclusion-learning procedure itself can be used as a tool to study other behavioral phenomena. In this symposium we present recent research that showcases the use of the exclusion-learning procedure, from basic studies to translational research to direct application. Catherine Graham (UNC-Wilmington) will present a study in which classes of equivalent stimuli were established through emergent exclusion performances. Richard Serna (UMass Lowell) will describe a study that examined factors that may limit the extent to which exposure to exclusion trials predicts accurate outcome performance. Deisy De Souza (Universidade Federal de So Carlos) will present research that extended exclusion-learning research with nouns to relations involving verbs and adjectives. Devon White (UMass Lowell) will present research that investigated the effects of using an identity-matching instead of arbitrary-matching baseline from which to introduce the exclusion protocol. William McIlvane (UMass Medical School) will serve as the symposium discussant.
|Keyword(s): conditional discrimination, emergent performance, exclusion learning, matching-to-sample|
|Can Stimulus Relations Established Only Through Exclusion Yield Equivalence?|
|CATHERINE ELIZABETH GRAHAM (University of North Carolina Wilmington), Carol Pilgrim (University of North Carolina Wilmington)|
|Abstract: Sidman (2000) has suggested that equivalence is a product of the reinforcement contingency, and that all positive elements of a contingency will become equivalence-class members. This study was designed to explore the possibility that stimuli never involved in a programmed reinforcement contingency could become members of an equivalence class. Typically developing 6 to 8-year-old children underwent AB and AC conditional discrimination training with class-specific reinforcers. Exclusion trials without programmed consequences were designed to establish XY relations, and YX symmetry trials were then presented. Identity matching was conducted next, with the X stimuli and the class-specific reinforcers from conditional discrimination training. Probe tests evaluated the formation of three equivalence classes (A1B1C1X1Y1, A2B2C2X2Y2, and A3B3C3X3Y3). Four of four participants demonstrated the exclusion phenomenon and YX symmetry relations. Two participants completed all training and testing phases and demonstrated the formation of 5-member classes, including the Y stimuli that were never related directly to reinforcers during training. This experiment documents the potential for stimuli never related directly to reinforcers to become equivalence-class members.|
|Limitations to Exclusion Learning: The Effects of Difficult-to-Discriminate Stimuli on Exclusion vs. Outcome Trials|
|RICHARD W. SERNA (University of Massachusetts Lowell), Michelle M. Foran (University of Massachusetts Lowell), Selena Tran (University of Massachusetts Lowell)|
|Abstract: This study examined factors that may limit the extent to which exposure to exclusion trials predicts accurate outcome performance. Specifically, we examined the effects of difficult-to-discriminate visual stimuli on both auditory visual exclusion performance and subsequent outcome performance in a matching-to-sample task. Participants were eight typically developing preschool children, ages 4-5 years old, all of whom entered the experiment capable of matching the spoken words “dog” and “cat” to line drawings of a dog and cat. This served as the exclusion baseline. The to-be-taught conditional discrimination (via the exclusion method) consisted of nonrepresentational forms that were very similar to one another, save for a single distinguishing feature, and the spoken nonsense words “veem” and “zid.” In an identity matching-to-sample pretest, four participants could match the nonrepresentational forms to one another and four could not. Three of the four that could not were trained successfully to do so with a stimulus-control shaping program designed to direct observing to the distinguishing stimuli. The participant acquired the discrimination without training. In subsequent exclusion-exposure tests, all participants showed highly accurate exclusion performance. However, in a test of conditional-discrimination outcome performance, the participants trained to discriminate the nonrepresentational forms failed to meet criterion, while three of the four participants who entered the study able to discriminate the forms met criterion. The results will be discussed in terms of the stimulus control engendered by exclusion trials and its interaction with observing behavior. Follow-up studies are ongoing.|
Probing Exclusion Responding With Auditory-Visual Conditional Discrimination Using Nouns, Adjectives, or Verbs as Sample Stimuli
|DEISY DAS GRACAS DE SOUZA (Universidade Federal de São Carlos), Thais Ribeiro (UFSCAR), Tamiris Gallano (UFSCAR)|
Responding by exclusion (fast mapping) is usually investigated using a baseline of auditory-visual conditional discriminations in which the auditory samples are names (i.e., word-object relations). Exclusion probes present a novel (undefined) noun and a novel stimulus displayed among defined comparisons; under a variety of conditions, participants typically select the unfamiliar comparison (emergent responding). Considering many other types of mapping relations within a language, this study investigated exclusion responding in relations involving verbs and adjectives. Typically developing children (24-29 months) participated in two conditions: Noun and Adjective (Experiment 1), or Noun and Verb (Experiment 2). The conditions were counterbalanced between participants. At baseline the auditory stimuli were Portuguese words equivalent to /Ball/, /Motorcycle/, /Airplane/ (Nouns); /Happy/, /Sad/, /Angry/ (Adjectives); and /Eat/, /Drink/, /Paint/ (Verbs). In exclusion probes samples were pseudo-words (respectively: /Fapi/, /Beva/; /Fob/, /Piva/; /Fobar/, /Mupir/); comparisons were pictures of unfamiliar objects (Noun), facial expressions (Adjective), or videotapes of actions (Verb). All participants showed exclusion in all three conditions; however, in the first trial scores for Adjectives and Verbs were lower than for Nouns (but above 75%), despite the fact that participants required more training to reach the baseline criterion in the Adjective and Verb conditions. The study replicated the exclusion responding with names, and extended it to other classes of words, thus suggesting that behavioral processes underlying fast mapping are the same (or highly similar). The exclusion procedure could be potentially useful in early intervention for children who are showing delays in language acquisition.
Using Identity Matching as a Baseline for Teaching Arbitrary Stimulus Relations With the Exclusion Method
|DEVON WHITE (University of Massachusetts Lowell), Richard W. Serna (University of Massachusetts Lowell)|
Learning by exclusion" is a teaching method used to rapidly establish new arbitrary stimulus relations. Typically, a baseline of known sample-comparison relations first is verified. Then exclusion trials are presented: a novel sample and S+ comparison and a known S- comparison from the baseline. Participants tend to exclude the S- in favor of selecting the S+. Outcome trials pit relations learned by exclusion against one another to verify that new conditional relations have emerged. A potential drawback to this method is that it requires a baseline of known arbitrary relations, which some individuals (e.g., with intellectual disabilities) may not be able to perform. This study examined whether an identity matching baseline -- often an easier task -- could replace an arbitrary baseline in a learning-by-exclusion protocol. Using a multiple-probe-across-participants design, three typically developing preschoolers were exposed to a computer-presented protocol of identity matching and exclusion trials, in which the known identity matching stimuli served as S- comparisons. Stimuli were pictures of fruit and single letters of the alphabet. All participants met criterion performance on outcome and maintenance trials. Future research should examine the necessary and sufficient conditions of the protocol, as well as its effectiveness with intellectually disabled individuals.