Association for Behavior Analysis International

The Association for Behavior Analysis International® (ABAI) is a nonprofit membership organization with the mission to contribute to the well-being of society by developing, enhancing, and supporting the growth and vitality of the science of behavior analysis through research, education, and practice.


36th Annual Convention; San Antonio, TX; 2010

Event Details

Previous Page


Poster Session #97
EAB Poster Session 1
Saturday, May 29, 2010
6:00 PM–7:30 PM
Exhibit Hall A (CC)
85. Foraging Behavior of Free Ranging Fox Squirrel Sciurus Niger
Area: EAB; Domain: Experimental Analysis
BRADY J. PHELPS (South Dakota State University)
Abstract: The visits of freely-roaming Fox Squirrels to two feeders will be videotaped. Two identical feeders approximately 24 inches apart in clear view from a video camcorder, programmed to record ten seconds of video once per minute, allowing time sampling of the squirrel’s foraging behavior. Placement of different foods such as highly preferred (shelled sunflower seeds, marketed as sunflower chips and hearts) relative to less preferred food (bird seed with some sunflower seeds interspersed) to non-preferred (safflower seeds). With different types of foods in the different feeders, I can attempt to quantify apparent food preferences over time. An additional intervention is planned: With feeders side-by-side, additional weights will be placed, in an incremental fashion over the course of several days, on the hinged lid of feeder with a preferred food. Each feeder will have a bolt inserted into the lid to allow the addition of weights to the bolt to progressively increase the effort required (the amount of weight that has to be lifted) to access a food; each weight is approximately 52 grams. The added weights will increase the effort expended to access the preferred food or drive the animal to choose the less- preferred food. It is predicted that much less weight/effort will be required to produce a shift in preference from a highly preferred food to a less preferred food than from a highly preferred food to a non-preferred food or from a less-preferred food to a non-preferred food.
86. Snake Learning: Acquisition and Extinction of an Appetitive Behavior by Brown Treesnakes (Boiga Irregularus)
Area: EAB; Domain: Experimental Analysis
JAMES DUENAS (University of Guam), Jesse Guerrero (University of Guam), Michael B. Ehlert (University of Guam)
Abstract: Extinction curves of operant behavior have been obtained from a wide variety of species since Skinner (1938) initially reported it. Most these studies used mammal and avian species. Reptile learning, however, is underrepresented in the literature, with investigation of appetitive snake behavior nonexistent. This presentation reports a study on behavioral extinction that used brown treesnakes (BTS), Boiga irregularus, the snake that decimated the avifauna of Guam. Tweleve wild-captured BTS learned to forage for carrion (frozen chicken meat and gecko) in a 3-meter x 3-meter foraging space. During Phase 1, eight task-naïve snakes served, while during Phase II, half the Phase I snakes were replaced by four new snakes. Using a free-operant arrangement, one day each week snakes gained access to prey by crawling along natural rope from a starting perch to four equidistant prey stations. Researchers baited all four prey stations during Phase I but baited only half during Phase II. Pre- and Post extinction procedure responses, as measured by time spent on a rope leading to a food station, were recorded and analyzed. Acquisition and extinction cures will be reported, along with a discussion on a laboratory model of snake learning.
87. Development of a Brief Paper-and-Pencil Measure of Temporal Discounting
Area: EAB; Domain: Experimental Analysis
CAITLIN M. PORTER (University of Central Arkansas), Shawn R. Charlton (University of Central Arkansas), Veda A. Charlton (University of Central Arkansas)
Abstract: Temporal discounting is a well-recognized, popular topic across a variety of research areas. Unfortunately, the rapid growth in the application of discounting has produced two key problems. First, a universally accepted, standard measure for conducting discounting studies is not available in the field. This lack of standardization causes difficulties for researchers trying to compare results across numerous studies and various populations. Second, as the applied implications of discounting become clearer, the need for a quick, easy-to-administer measure of discounting will increase. This poster presents the product of a series of studies aimed at refining the 27-item discounting inventory, created by Kirby and Markovic (1996), into a measure of discounting that can be easily administered, quickly scored, and effectively applied, across a variety of research questions/contexts. The validity and reliability of the measure will be discussed as well as its ability to replicate key findings in the discounting literature such as the commodity effect and the magnitude effect.
88. Some Effects of Varied Delays During Training on Delayed Matching-to-Sample in Humans
Area: EAB; Domain: Applied Behavior Analysis
Katherine A. Ericsen (Allegheny College), Adam M. Smith (University of Pittsburgh), RODNEY D. CLARK (Allegheny College)
Abstract: The general purpose of this study is to replicate and extend upon research reported by Sargisson and White (2001). Sargisson and White’s article titled Generalization of Delayed Matching-to-Sample Following Training at Different Delays demonstrated how accuracy in matching behavior with simple stimuli in pigeons is not necessarily only affected by latency, but is affected by discrimination at the time of remembering (as cited in Pierce & Cheney, 2004). Thus, if a pigeon is trained to match colors using a 4 s delay, the birds will match more accurately at that delay than at delays that are shorter and/or longer than 4 s. The present experiment tests this phenomenon with humans as subjects and compares matching of simple stimuli (15, 30, and 45 sec delays). Thus, the specific purpose of the following study is to determine whether or not matching ability is affected by the temporal distance at which humans are trained to match and whether or not there is a difference in accuracy levels when matching simple stimuli versus complex stimuli under different training delays. When subjects were trained at the 15sec. delay fewer errors occurred then at the other delays. When training was set at 30 sec., again, fewer errors occurred than at the other delays. Finally, when training was set at 45 sec., no differences were observed among any of the delays.
89. The Effect of an Collaboration Intervention Strategy of the Videotape-Based Introspection and Token Reinforcement Methods on the Behavioral Problems of ADHD Children
Area: EAB; Domain: Experimental Analysis
JUNG YEON CHO (Daegu Cyber University), Shin-Hee Kim (Daegu Cyber University)
Abstract: This study aims at examining the effects of an integrated intervention of the videotape-based introspection and token reinforcement methods on the behavioral problems of ADHD children, such as leaving their seats, inattentive and disruptive behaviors. In this study, we chose as a study subject a child who attends preschool affiliated with P Elementary School located in Jeonju. The child was diagnosed with ADHD since he fell into 6 items of inattentiveness and 9 items of hyperactivity and impulsivity according to ADHD criteria from the DSM-?(1994), and scored 17 points on attention (6 items), 24 points on hyperactivity (5 items), 9 points on hostile behavior (6 items), and 10 points on social skills, whose sum fell below 25 percent of the total score on the ADD-H Comprehensive Teacher’s Rating Scale (ACTeRS). This study adopted the multiple-baseline across behaviors for a child to examine the effects of the videotape-based introspection and token reinforcement methods on the ADHD child’s behavioral problems. It consisted of three phases, namely, baseline, intervention 1, intervention 2, and maintenance, regarding three behavioral problems, leaving a seat, inattentive behavior, and disruptive behavior, trying to prove the effects of these interventions to improve behavioral problems and learning behavior. The results of the ACTeRS conducted before and after the videotape-based introspection and token-reinforcement intervention showed that the child showed attention (10 points), hyperactivity (24 points), social skills (10 points), and hostile behavior (21 points) before the intervention, and after interventions 1 and 2, attention (19 points), hyperactivity (15 points), social skills (15 points), and hostile behavior (14 points), indicating a positive change in his/her behavioral problems.
90. Other Behaviors Influencing the Occurrence of Resurgence for Pigeons
Area: EAB; Domain: Experimental Analysis
SATOSHI OBATA (Tokiwa University), Tetsumi Moriyama (Tokiwa University)
Abstract: Resurgence is defined as the reoccurrence of a previously reinforced behavior under the condition that delivery of reinforcer ceases for a more recently reinforced behavior. Lieving and Lattal (2003) conducted experiments with pigeons to assess the experimental conditions necessary for the occurrence of resurgence. Although they showed some variables controlling resurgence, they did not investigate the relationship between the behavior to be recurred and other behaviors. It may be that the occurrence of other behaviors influences on the resurgence. We investigated this issue with pigeons using similar experimental procedures to those of Lieving and Lattal. Five pigeons were used as the subjects. They were trained to peck the key of the standard operant chamber. After that their key-peck responses were extinguished and then their treadle pressing were reinforced. Finally their treadle presses were extinguished to investigate whether the resurgence of key pecking occurred. Only two pigeons showed the occurrence of resurgence. They also emitted various behaviors relating to their key-peck behaviors. These results suggest the possibility of the occurrence of other behaviors as a controlling variable of the resurgence for pigeons.
91. Using an Adjusting Amount Procedure to Investigate Impulsivity in Spontaneously Hypertensive Rats (SHR)
Area: EAB; Domain: Experimental Analysis
P. A. HALSEY (James Madison Univeristy), Sherry L. Serdikoff (James Madison University)
Abstract: This experiment compares impulsivity in spontaneously hypertensive rats (SHRs), a putative animal model of Attention Deficit/Hyperactivity Disorder (ADHD), with two control strains. . Impulsive behavior has been operationally defined as preference for smaller sooner (SS) reinforcers over larger later (LL) reinforcers when both are made concurrently available in the context of discrete trial choice procedures. In this experiment, the LL reinforcers vary across 10 conditions (5 and 10 pellets, each evaluated at five delays (2, 4, 8, 16, and 32 s). Within each LL condition, the amount of the SS reinforcer starts at 1 pellet delivered after .5 s and the amount is adjusted for each rat based on its own choices in order to identify an indifference point for that condition.. The resulting data are fit to the hyperbolic-decay model to determine the extent to which the delayed reinforcers are discounted. When an individual is more impulsive, the discounting function is steeper. We evaluate the extent to which the SHRs show more discounting than control rats in this procedure in the context of recent debates regarding the adequacy of the SHR animal model of ADHD.
92. Equivalence Training in the Rat: Effect of Asymptomatic Training of Simple Relations
Area: EAB; Domain: Experimental Analysis
ERICA ALEJANDRA BERTEL FERREIRA (Fundacion Universitaria Konrad Lorenz), Edth Leal (Fundacion Universitaria Konrad Lorenz), Angelica Maria Osorio (Fundaction Universitaria Kondrad Lorenz), Cristina I. Vargas-Irwin (Fundación Universitaria Konrad Lorenz)
Abstract: We evaluated the effect of variations of arbitrary relations training on emergent relations in a conditional discrimination task in 8 rats. The subjects were exposed to training in four conditional discrimination relations. For each of these relations, four tests of symmetry and two transitivity tests were performed. Each subject started with the training of a relationship, met the criterion or response requirement, once these requirements were achieved the the symmetry of the relationship trained was tested. This procedure was conducted for the four relationships. The two training criteria used were are 95% of responses in two continuous sessions for one group of animals, whereas a second group hat to reach 70% of correct trials for two consecutive sessions. Transitivity was also tested for each pair of conditional relationships. The differential effects of the two training criteria on the symmetry and transitivity test are reported.
93. Biobehavioral Service Outpatient Clinic: A Description of Hypothesis Driven Clinical Process
Area: EAB; Domain: Applied Behavior Analysis
SHEEHAN D. FISHER (University of Iowa), David P. Wacker (University of Iowa), Jeffrey R. Luke (University of Iowa), Maliha Zaman (University of Iowa), Julianne Elizabeth St John (University of Iowa), Yaniz C. Padilla Dalmau (University of Iowa)
Abstract: The purpose of this poster is to provide a detailed description of the clinical model used at the Biobehavioral Service Outpatient Clinic of the University of Iowa. The Biobehavioral Service sees children and adults with developmental disabilities (e.g., autism, mental retardation) who display problem behavior (e.g., self-injury, aggression, destruction of property) during brief 90-120 minute clinic visit. A collaboration of psychologists, speech-language pathologists, social workers, medicine, and other pediatric professionals assist in developing an assessment and treatment plan based on the particular needs of the client and the referral concerns. The Biobehavioral clinic team uses a hypothesis driven model to develop assessment and intervention evaluations. Case examples are presented in order to illustrate the clinical process used to develop hypotheses, to assess the functions of the clients’ behaviors and to develop clinical interventions that can be implemented by the clients’ caretakers and community services (e.g., school, social service). Current applications of brief functional analysis procedures will be highlighted. Trained coders evaluate the assessment and treatment of each case and have good interrater reliability (i.e. 90% or greater).
94. Analysis of the Different Patterns of Exercise Behavior Shown by Hemodialysis Patients
Area: EAB; Domain: Applied Behavior Analysis
ITOKO TOBITA (Osaka Jikei Research Center of Health Care Management), Sumie Suzuki (Dokkyo Medical University), Masato Ito (Osaka City University)
Abstract: Study Objective: Analysis of the different patterns of exercise behavior shown by hemodialysis patients. Participants: A total of 30 hemodialysis patients were assigned to the intervention and control group: those in the former participated in a behavior modification program in addition to the exercise program, whereas those in the latter participated only in the exercise program. This study was approved by the Ethics Committee of the Osaka University. Intervention: The behavior modification program included a joint agreement on the exercise strength; this was a part of the introduction phase of the program. Graphic feedback, verbal reinforcement, and a sticker were provided in phase one. The detailed benefits of the exercise program were explained in phase two. Reinforcement schedules: The instruction phase spanned over the first 3 weeks, VR (3 times/week) during weeks 4-10, and VR (1-2 times/week) during weeks 11 to 16, extinction was carried out during weeks 17-20 and VR (1-2 times/week) during weeks 21-24. Results: Three patterns of exercise behavior were identified: continuous, non-continuous, and unstable patterns. The numbers of participants showing these patterns in the intervention group were 12, 0, and 2, respectively; the corresponding numbers in the control group were 7, 6, and 3.
95. Can Conditional Stimuli Function as Conditioned Reinforcers for Simple Discrimination Acquisition and Equivalence-Class Expansion?
Area: EAB; Domain: Experimental Analysis
WILLIAM HOGAN (University of North Carolina, Wilmington), Kristin Wilkinson-Yonkers (University of North Carolina, Wilmington), Alicia Rae McLamb (University of North Carolina, Wilmington), Jennifer Irene Stuart (University of North Carolina, Wilmington), Carol Pilgrim (University of North Carolina, Wilmington)
Abstract: In the present study, six 4 to 10 year old typically developing children learned simple discriminations with class-specific reinforcers, such that choosing B1 produced Reinforcer 1, choosing B2 produced Reinforcer 2, etc. Emergent conditional discriminations between the B stimuli and reinforcer images were then demonstrated on probe trials. Simple discriminations were then trained with A stimuli and reinforcer probes again revealed emergent relations, as did AB conditional discrimination probes. Next, a CD conditional discrimination was trained using the same class-specific reinforcers. Subsequent phases will determine whether the C stimuli used as conditional stimuli in one phase of the experiment will function as conditioned reinforcers in a later phase of the experiment, thus facilitating equivalence-class development and expansion. The results of this study thus far support Sidman’s (2000) theory that all members of equivalence classes, from sample stimuli to class-specific reinforcers, can become truly equivalent to one another.
96. The Effects of Response-Cost Punishment on Rule Following in a Choice Task
Area: EAB; Domain: Experimental Analysis
ADAM E. FOX (Western Michigan University), Cynthia J. Pietras (Western Michigan University)
Abstract: An Important variable controlling human behavior is verbal rules or instructions. The present study is investigating how penalties for breaking rules affects rule following. Participants were presented with choices between progressive and fixed-time schedules of reinforcement and were given instructions (rules) for how to respond to maximize earnings. Across sessions, the progressive-schedule step size was manipulated so that the rules became increasingly inaccurate. In some conditions deviating from the instructions produced money losses (response-cost penalties). Participants were exposed to penalty and no-penalty conditions in a counterbalanced order to determine how a history of punishment for rule-breaking influeneces subsequent rule following. Preliminary results indicate large individual differences in the extent of control by the rules, and that the penalty has little additional affect on choice.
97. Effects of Variable Interresponse Time on Humans’ Reinforcement Schedule Sensitivity
Area: EAB; Domain: Experimental Analysis
NAOKI YAMAGISHI (Ryutsu Keizai University)
Abstract: The author examined whether the variable interresponse time (IRT) increase humans' sensitivity to reinforcement schedules. One of two groups of participants was exposed to test phases before and after the training phase. In the test phase, sensitivity to reinforcement schedule was measured. In the training phase, lag schedule for IRT was used to increase IRT variability. This reinforcement schedule was designed to give 10 points for IRT that is different from preceding one. Criterion of the difference for reinforcement was 1 second and more as absolute value. Another group was exposed to test phase twice only. The results indicate that participants were more sensitive to the reinforcement schedules after the training phase. These findings are discussed in terms of functioning of IRT differentiation by the generation of variable IRT, and sensitivity generated by the IRT differentiation. It was suggested that behavioral variability is one of most important source of humans’ sensitivity to reinforcement schedules.
98. Probability Discounting: Does Age Affect Risk Sensitivity?
Area: EAB; Domain: Experimental Analysis
MANISH GOYAL (Western Michigan University), Gabriel D. Searcy (Western Michigan University), Tobey L. Schipper (Western Michigan University), Cynthia J. Pietras (Western Michigan University)
Abstract: The study was designed to systematically replicate a study by Green, Fry, and Myerson (1994) that found that children were more impulsive than young adults and older adults on a delay-discounting procedure. All participants chose between small certain and larger but probabilistic hypothetical monetary amounts. The probability of the large amount was varied across blocks of trials. Two large amounts were investigated. At each probability the amount of the small certain option was increased across trials to determine an indifference point between the two options. Eleven children, 17 young adults and 15 older adults showed orderly probability discounting. Results showed that the children and young adults discounted the smaller reward at a much slower rate than the larger reward indicating greater risk proneness for the smaller reward while the older adults showed the opposite. The rates of discounting were much higher for the older adults for both amounts compared to the other two groups. The young adults showed lower discounting rates than children. The results are consistent with Green, et al. showing that there is an age related difference in probability discounting, with the older adults showing more probability discounting (risk aversion) than younger adults or children.
99. Recovery of Maintained Generalization Gradient Form by Pigeons
Area: EAB; Domain: Experimental Analysis
JAMES CERRI (University of Tennessee), John C. Malone (University of Tennessee)
Abstract: Maintained generalization gradient forms produced by 3-day averaged response rates were examined to determine to what extent previously-learned response patterns persist after a change in reinforcement contingencies. White Carneau Pigeons were exposed over 34 months to two different sequences of quasi-random varied frequencies of light flicker ranging between 13 and 37-Hz. The first sequence consisted of S+ Center Condition where 25-Hz flicker was paired with VI reinforcement and all other stimuli were presented in extinction, soon producing an inverted V-shaped gradient form. Subjects were then presented with S+ Extremes Condition where the two high- and low-extreme frequencies were presented with VI reinforcement, producing a U-shaped gradient form. These conditions appeared in an AB, ABA, ABA order. Specific results included asymmetrical forms after symmetric forms had been previously produced and “W” forms during initial sessions directly after switching conditions. After initial training periods, the S+ Centers Condition gradient form was not recoverable after the interposing of the S+ Extremes Condition. For three of four subjects during the final presentation of S+ Extremes, recovery of gradient form occurred within 9 sessions and during the final presentation of S+ Center recovery of gradient form occurred within 12 sessions.
100. Reciprocity of Responding and Induction of Verbal Exchange as Determinants of Partial-Altruistic Behavior in Humans
Area: EAB; Domain: Experimental Analysis
Emilio Ribes Iñesta (Universidad Autonoma de Guadalajara), NORA RANGEL (Universidad Autonoma de Guadalajara), Lizbeth Pulido Avalos (Universidad Autonoma de Guadalajara), Oscar Vazquez (Universidad Autonoma de Guadalajara), Hugo Reyes (Universidad Autonoma de Guadalajara), Elizabeth Ramirez (Universidad Autonoma de Guadalajara)
Abstract: Eight university students were randomly distributed in eight dyads (each participant had a confederate peer in the dyad). Participants were not informed that their peer was an experimenter’s confederate. In a situation of partial altruism, dyads solved a visual puzzle on two synchronized computers screens. Participants and confederates could track the performance of his/her peer, and place pieces in either puzzle. A within-subject design was used, comprising two individual baselines and five experimental phases, each one involving a specific confederate’s behavior toward a peer in the dyad: the percent of reciprocal placing of pieces in the peer’s puzzle. Dyads were randomly distributed in two different groups. Dyads in Group 1 were exposed to an ascending order of the percentage of reciprocity by the confederate (0, 25, 50, 75, 100%), and dyads in Group 2 were exposed to a descending order (100, 75, 50, 25, 0%). At the beginning of each session, in both groups, confederates induced verbal exchange in participants through a written questionnaire. Results are discussed in terms of the effects of induction of verbal exchange and the percentage of reciprocity by the confederate in social partial-altruism contingencies.
101. The Role of the Existing Repertoire in the Generation of Novel Behavior (2)
Area: EAB; Domain: Experimental Analysis
APRIL M. BECKER (University of North Texas), Jesus Rosales-Ruiz (University of North Texas), Brett Grant Kellerstedt (Salem State College)
Abstract: The purpose of this experiment is to show the role of the existing or newly created repertoire in the generation of novel topographies and to replicate the results of a similar experiment in this lab. Rats were exposed to a variation contingency (VAR) wherein topographically varied physical contact with an object was consequated with food. The object was one of four; objects were cycled sequentially so that the rat experienced a one-minute session with each before it was exchanged for the next. The rate of emission of variable responses increased and then stabilized for all objects, while the emission of completely novel responses (those never before emitted in the experiment) increased at first but decreased to zero over time. A novel behavioral “atom” was then trained (IV) for each object at different times using a shaping procedure in a multiple baseline design. The VAR condition was re-imposed to see if this newly trained behavior would be emitted or combined with other responses to produce novel responses (DV) as a function of the new training. This research may suggest strategies for inducing creativity and may inform our understanding of the variables that determine the emission of new behavior. Results are pending.
102. Increasing Independent Eating Using Avoidance Conditioning
Area: EAB; Domain: Applied Behavior Analysis
G. JOSEPH SCHLERETH (Kennedy Krieger Institute), John C. Borrero (University of Maryland, Baltimore County), Rinita B Laud (Kennedy Krieger Institute), Carrie S. W. Borrero (Kennedy Krieger Institute)
Abstract: Avoidance conditioning procedures maintain responding through negative reinforcement. The purpose of this study was to evaluate the efficacy of avoidance in the maintenance of independent eating for a 4-year-old male admitted to an intensive program for children with feeding disorders. First, compliance with eating a developmentally appropriate volume was established using escape extinction. Second, independent eating skills were evaluated under a baseline condition in which prompts to eat were not given, escape extinction was withdrawn, and the meal was terminated if the child failed to take a bite or drink after 3 min. Independent eating was then evaluated during a conditioned avoidance procedure in which uninterrupted access to a preferred video was provided contingent on independent eating at least once every 30 s. If independent eating was not observed, access to the video was interrupted by turning the television off and on every 10 s. Results indicate that the baseline rates of independent eating were lower than rates observed during treatment. Higher rates of independent eating were observed as the child learned to eat faster to avoid interrupted access. Parallels to basic research and use of the procedure as a post-escape extinction treatment are discussed.
103. Effect of Three Kinds of Training on a Second-Order Matching-to-Sample Task
Area: EAB; Domain: Experimental Analysis
EMANUEL MERAZ MEZA (Universidad Veracruzana), Augustin Daniel Gomez Fuentes (Universidad Veracruzana), Enrique Zepeta Grcia (University of Veracruz), Cecilia Magdalena Molina Lpez (University of Veracruz), Mario Serrano (Universidad Veracruzana)
Abstract: An experiment was done to evaluate the effect of the complexity of three types of descriptions that included texts concerning instances, modalities and relations in a correspondence training (saying – doing – describing). The experimental task consisted of a second-order matching-to-sample situation. Involved three groups of five universities in each group, they were between 18 and 23 years old, experimentally naive. The results showed low scores in group 3, which was used texts of relations. The results are discussed in terms of linguistic interactions in discrimination procedures.
104. Contextual Variables Interfere With Exclusion of Novel Names and Verbs
Area: EAB; Domain: Experimental Analysis
Aline Roberta Aceituno Costa (Universidade Federal de São Carlos), DEISY G. DE SOUZA (Universidade Federal de São Carlos), Julio C. De Rose (Universidade Federal de São Carlos)
Abstract: Exclusion responding has been extensively documented with name-object relations: When an undefined name is dictated, participants often select an undefined object or picture displayed as a comparison stimulus. The present study sought to identify conditions under which participants would not select an undefined stimulus. Pseudo-words simulating names or verbs were presented as sample stimuli on exclusion probes. When the baseline presented pictures of objects (and the mask) as comparison stimuli, the presentation of a pseudo-name on exclusion probes always resulted in the selection of an undefined picture; however, when the sample was a pseudo-verb participants often selected the mask, rejecting both, the defined and the undefined pictures. The opposite was observed when the baseline presented video clips of defined actions as comparisons and the sample was a dictated verb: The presentation of a pseudo-verb always occasioned the selection of the video clip of an undefined action, but a mask was selected when the sample was a pseudo-name. These data suggest that the discrepancy between the classes of sample and comparison stimuli interferes with the exclusion responding and support the notion that the formation of a stimulus class involving simultaneously samples and comparisons is a necessary condition for exclusion.
105. A Connectionist Model of Stimulus Class Formation Using a Yes/No Procedure and Compound Stimuli
Area: EAB; Domain: Experimental Analysis
ANGEL TOVAR Y ROMO (Universidad Nacional Autónoma de México), Alvaro Torres (Universidad Nacional Autónoma de México)
Abstract: We analyze the stimulus class formation in an artificial neural network (ANN) and in a human experimental study with a “YES/NO” procedure; this procedure is an alternative to the matching to sample procedures for the analysis of equivalence class formation. For the connectionist simulation a feed-forward back-propagation ANN was designed with 9 inputs, 4 hidden and 2 output nodes. During the training phase, the conditional relations A1B1, A2B2, B1C1 and B2C2 were established for “YES” response, while A1B2, A2B1, B1C2 and B2C1 were established for “NO” response. During the test phase we probe the emergence of A1C1 and A2C2 for "YES" response, and A1C2 and A2C1 for “NO" response. A similar procedure was carried out with six female participants aged 19-22 years. The conditional relations of the training and testing phases were similar in the ANN and in the human experiment. The results showed that both the ANN and the human participants could respond with the established class membership. We discuss the efficacy of the YES/NO procedure for simulating the stimulus class formation in a connectionist model in contrast with the models inspired in matching to sample procedures.



Back to Top
Modifed by Eddie Soh