The Huron University College Journal of Learning and Motivation Volume 51 Issue 1 Article 5 2013 Resistance to Impulsivity and Temporal Discounting in Canis lupus familiaris Willey Dow Follow this and additional works at: http://ir.lib.uwo.ca/hucjlm Part of the Psychology Commons Recommended Citation Dow, Willey (2013) "Resistance to Impulsivity and Temporal Discounting in Canis lupus familiaris," The Huron University College Journal of Learning and Motivation: Vol. 51: Iss. 1, Article 5. Available at: http://ir.lib.uwo.ca/hucjlm/vol51/iss1/5 This Article is brought to you for free and open access by the Psychology at Scholarship@Western. It has been accepted for inclusion in The Huron University College Journal of Learning and Motivation by an authorized administrator of Scholarship@Western. For more information, please contact kmarsha1@uwo.ca.
Resistance to Impulsivity and Temporal Discounting in Canis lupus familiaris Willey Dow Huron University College Based on the Petter, Musolino, Roberts, and Cole (2009) finding that dogs were able to utilize cues to choose a baited container at a rate significantly higher than they chose an empty container and the Cole (1990) finding that albino rats continued to press a lever after an initial lever press delivered a foot pellet when additional lever presses banked pellets the current study predicted that a dog could learn to make decisions that resist impulsivity for a reward delivered after a short temporal delay. The results of this experiment support that prediction. The subject chose to consume a smaller quantity of food from a white container significantly more often than an equal or larger quantity of food from a black container when rewarded for doing so. The decision making capabilities of dogs (Canis lupus familiaris) and their ability to resist impulsivity has long been debated by pet owners. It has also been experimentally investigated by university researchers. Petter, Musolino, Roberts, and Cole (2009) found that dogs were able to utilize cues to choose a baited container at a rate significantly higher than they chose an empty container. The cues utilized were a cooperative person pointing to a baited container vs. a deceptive person pointing to an empty container or a white baited container vs. a black empty container. Therefore with assistance from either human or environmental cues dogs do appear capable of making rewarding decisions. Cole (1990) found that albino rats (Rattus norvegicus) of the Wistar strain continued to press a lever after an initial lever press delivered a foot pellet when additional lever presses banked pellets in a fashion resembling a savings account or when banked pellets earned interest. Rats were able to resist consuming an available
food pellet in favour of performing a behaviour that would deliver a larger reward after a temporal delay. Therefore a rat can learn to resist impulsivity. This research questions whether a dog is capable making decisions that resist impulsivity for a reward delivered after a short temporal delay. Can a dog learn to choose a smaller reward now in favour of a larger reward later? Can a dog make a decision to consume a very small quantity of moderately desirable food over a much larger quantity of the food when later rewarded for doing so with a more highly desirable food? If a dog can resist the larger immediate reward for a delayed reward it will support the current experimental hypothesis that a dog can learn to make decisions that resist impulsivity and temporal discounting. Method Subject The subject was a female dog {Canis lupus familiaris) of the West Highland White Terrier breed purchased at 6 weeks of age from a local breeder. She was given the name Frosty and kept as a family pet in the experimenter's home. At the time of experimentation the dog was 9 years old. During her last yearly checkup at Yates Veterinary Hospital in Woodstock her weight was 7.71 kg. This year's checkup took place 1 week following the completion of experimentation. Her weight was 7.48 kg and she was given a clean bill of health at that time. The dog's daily diet was 75 g of Now Fresh Grain Free Senior Dog Food Recipe (NF) produced by Peteurean in Canada given at approximately 9am and 5pm daily. Water was always provided ad lib. Both food and water were delivered to a pair of stainless steel bowls dedicated to that use. The dog's diet was also supplemented with human food stuffs and a variety of commercially available dog treats. T Bonz Porterhouse Flavour Treats (TB) produced by Nestle Purina
Petcare in Mississauga had elicited a particularly enthusiastic response from the dog in the past and were utilized in the current research. Experimental sessions took place during normal feeding times. The total amount of food that was given during typical feedings was available to be consumed during experimentation. This approximated a closed food economy but any NF not consumed during experimentation was provided to the dog immediately after each session. Every effort was made to minimize any disruption to the dog's diet by feeding it more or less than it was accustomed to eating. To do so the quantity of TB consumed during each session was calculated and the amount of NF given following each session was adjusted based on that calculation. The dog has had no experimental history per se at the time of the current research. She has however been trained to perform behaviours using the principles of operant conditioning for the convenience and amusement of her owners Graham and Alice Dow. She has been trained to indicate a need to go outside or to come back inside by alerting her owners by voice and by then proceeding to the back door. She has also been trained to "sit", to lower her hind quarters to the ground, to "stay", to remain in that position and to "come", to proceed to the location of her owner. The dog's ability to sit, stay and come when instructed where relied upon during the current research. Apparatus The experimental apparatus consisted of a hallway within my home and two overturned stainless steel dog dishes. Figure 1 depicts the layout and measurements of the experimental apparatus as well as the subject's perspective. One end of the hallway opens to a kitchen and the border between the kitchen's tile floor and the hallway's hardwood floor represented the start line. The opposite end of the hallway has 4 rooms and a closet, the doors to which remained closed during trials. The overturned dishes
Figure 1. Orientation of experimental apparatus from an overhead perspective and from the subject's perspective. Overhead PCTspective Subject Perspective 86cm Start Conrect choice ri^t
were placed near the closed end of the hallway. From the subject's perspective one dish was placed on the right side and one on the left side, the centre of each being 272 cm from the start line. The exterior surfaces of one dish were covered with white paper (correct choice), a single piece of NF was placed on that paper cover and a quantity of TB placed beneath. The exterior surfaces of the other dish were covered in black paper (ineorreet choice), a quantity of NF was placed on that paper cover and a quantity of TB was fixed to the interior surface of the dish in order to remain inaccessible and to control for scent. The relative positioning of the dishes, left or right was randomized using a random number generator to control for the development of a directional preference. The positioning of the first trial in each session was also alternated to control for the possibility of developing a directional preference based on that first trial. An adjoining room was used to hold the subject between trials. Procedure During each trial the subject was placed behind the starting line. The experimenter then verbally instructed the subject to "sit" and "stay", visually emphasizing this by placing an open palm in front of her face. The subject remains seated while the experimenter walks away from the subject and past the white and black overturned dishes. The experimenter then turns to face the subject and verbally instructs her to "come". The subject then proceeds towards the experimenter andfindsthe two dishes. Each overturned dish has NF on top of it and when the subject consumes the NF from either dish that will constitute the subject's choice and that dish will then be lifted by the experimenter. Figure 2 depicts the quantities of NF on top of each overturned dish and the quantity of TB to be found beneath the correct choice across all phases of this experiment. If the subject made the correct choice she was permitted to consume the TB
Figure 2. Quantities of Now Fresh Grain Free Senior Dog Food Reeipe (NF) representing choices and T Bonz Porterhouse Flavour Treats (TB) representing rewards across all phases of training and experimentation. Whitecorreet Blackincorreet Training trials 1-40 Experimental trials 1-20 Experimental trials 21-30 Experimental trials 31-40 Experimental trials 41-45 0.36g 0.36g 0.36g 0.36g 0.36g NF/3.38g NF/3.38 TB NF/13.5gTB NF/13.5g TB NF/13.5g TB TB 0.36gNF/- 3.6gNF/- 13.32g NF/- 13.32g NF/- 13.32 gnf/-
found beneath by lifting that dish. If she made the ineorreet choiee she would be shown that there was no accessible TB by lifting that dish. The subject was placed in an adjoining room between trials. During training trials 1 to 40 and experimental trials 1 to 30 the experimenter assisted the subject to make the correet choice by pointing to it while instrueting her to come and by praising her verbally and affectionately following eorreet choices. During experimental trials 31 to 45 no experimenter assistance or praise was used. Results Figure 3 displays the frequency of correct and incorrect choices across all phases of experimentation. During training trials there was a statistically significant preference for the eorreet choice x,^(l, V= 40) = 5,p <.05. The subject chose the correct choice on 75% of these trials. During experimental phases 1 and 2 there was a statistieally signifieant preferenee for the correct choice x^(l, V = 30) = 6.67, p<.05. The subjeet ehose the eorreet ehoiee on 83% of these trials. During experimental phases 3 and 4 there was a statistically significant preference for the correct choice x^(l, V= 15) = 5.63, p <.05. The subject chose the eorreet ehoiee on 93% of these trials. Diseussion This researeh questioned whether a dog is eapable making deeisions that resist impulsivity for a reward delivered after a short temporal delay. If the dog could resist the larger immediate reward for a delayed reward it would support the current experimental hypothesis that a dog ean learn to make decisions that resist impulsivity and temporal diseounting. As expected the correct choice was made at a rate significantly higher than chance under all conditions analyzed.
Figure 3. Subject's choices over all phases of training and experimentation. Phase Choice Frequency Training eorreet 30 trials 1-40 ineorreet 10 Experimental Phase 1 correct 14 trials 1-20 incorrect 6 Experimentid Phase 2 correct 9 trials 21-30 incorrect 1 Experimental Phase 3 correct 10 trials 31-40 incorrect 0 Experimental Phase 4 correct 4 trials 41-45 incorrect 1
For the purposes of the statistical analysis the training trials were analyzed together because all training trials were eondueted in the same maimer. The subject was faced with an equal quantity of NF correet ehoiee and NF ineorreet ehoiee while the experimenter assisted by pointing to the correct choice during eaeh trial and by praising eorreet choices when made. Experimental phase 1 and experimental phase 2 were analyzed together because they were similar in that the subjeet was faced with an increased quantity of NF incorrect choice, increasing the incentive of making an ineorreet choice while the experimenter assisted by pointing to the correct choice in eaeh trial and by praising correct choices when made. Experimental phase 3 and experimental phase 4 were analyzed together because they were similar in that the subjeet was faced with an increased quantity of NF incorrect choice, increasing the incentive to make the incorrect choice in each trial but without the element of experimenter assistance or praise used during previous phases. The progression from training through experimentation was designed as such to resemble a parent teaching a child resistance to impulsivity. The training phase approximates an environment existing during early childhood. Parents have the greatest control during this period. They still have their child's attention and are still largely able to control their environment and the rewards they ean obtain within that environment. As children age they are faced with increasingly attractive incentives for making the incorrect choice and morefreedomto do so, just as the subject in the present study had been. As expected the eorreet choiee was made at a rate significantly higher than chance imder all conditions analyzed. The support was encouraging but would have been more so with use of an augmented experimental procedure. The procedure utilized
included having the experimenter walk away from subject, proceeding past the subject's choices while she remained at the starting line. This involved relying on the subject's ability to sit, stay and come when instructed. It also involved the experimenter remaining visible to the subject during trials. Every effort was made to standardize the experimenters actions from trial to trial as to not affect their outcomes but it would have been preferential to eliminate that possibility altogether. Perhaps it would be preferable for the experimenter to remain at the starting line, having the subject on a leash, releasing her to begin a trial, pulling her back to the starting line after each choice, and rewarding her there if she had made the correct choiee. This would however eliminate the possibility of assisting the subject by pointing to the correct choice. It would also eliminate the need to rely on the dog's ability to sit, stay and come which can be troublesome. There were therefore advantages to each procedure. Ultimately the decision was made in favour of permitting experimenter assistance and praise during trials. The exclusive use of the experimenter's family pet may also have contributed to the extent to which the results were supportive of the current studies hypothesis. Dogs have been bred for thousands of years to be loyal to humans. This dog has that common inheritance and has lived her entire 9 year life with the experimenter. She may have more motivation to please the experimenter than to consume the food available on the ineorreet ehoiee or to consume the reward for making the correet choice. Based on this line of reasoning it may be advisable to continue with this line of experimentation with the aforementioned procedural improvement but with no food reward for correct choices. Would a subject learn to resist a food reward in return for experimenter praise alone? This experimenter expects that a canine subject could be taught to resist
impulsivity in this way and that such an experiment would return significant results if initial training involved a food reward. Also based on this line of reasoning it may be advisable to replicate this experiment with a dog or dogs that are unfamiliar to the experimenter. This could mitigate both the potential effect of a long standing relationship and the potential for experimenter expectations affecting the choices made during trials. It is also possible that the subject has a preference for white as opposed to black dishes. The placement of the correct choice was randomized but the eorreet choice remained the white dish on all trials. If this experiment were to be replicated with the aforementioned improvements it would be advisable to present the alternate configuration, black choice is correct, white choice is incorrect to half of the subjects to control for this possibility. The results of the current study were supportive of the experimental hypothesis that a dog is eapable making decisions that resist impulsivity for a reward delivered after a short temporal delay. The correct choice was made at a rate significantly higher than chance under all conditions analyzed. Improving the procedural elements described above could enhance confidence in those results. The results do suggest that resistance to impulsivity should be taught early in a child's life when that child's environment is controlled by parents to the greatest degree. Children should also gradually be made aware of the short term rewards that an incorrect decision would provide. They will need to be inoculated against their inevitable presence later in life.
References Cole, M. R. (1990). Operant hoarding: A new paradigm for the study of self-eontrol. Journal of the Experimental Analysis of Behavior, 53, 247-261. Petter, M., Musolino, E., Roberts, W. A., & Cole, M. (2009). Can dogs (Canis familiaris) detect human deception? Behavioural Processes, 82, 109-118.
Appendix A Raw Data Appendices
Appendix A Training Phase Now Fresh Grain Free Senior Dog Food Reeipe (NF) and T Bonz Porterhouse Flavour Treats (TB) Session 1 Trials 1-20 Correct choice (white) is 1 piece NF (0.36g) rewarded with % pieee TB (3.38g), ineorreet ehoiee (blaek) is 1 piece NF (0.36g), experimenter assisted by pointing to correct choice and by praising correct choice Session 2 Trials 21-40 Correet ehoiee (white) is 1 pieee NF (0.36g) rewarded with % piece TB (3.38g), incorrect choice (black) is 1 piece NF (0.36g), experimenter assisted by pointing to correet ehoiee and by praising eorreet ehoiee Trial # Correet Correet Ineorreet Trial # Correet Correct Incorrect Choiee Choiee Choice Choice Choice Choice Location Loeation 1 Left V 21 Right V 2 Right V 22 Left V 3 Right V 23 Right V 4 Left V 24 Left V 5 Left V 25 Left V 6 Right V 26 Right V 7 Left V 27 Left V 8 Right V 28 Right V 9 Right V 29 Right V 10 Left V 30 Left V 11 Left V 31 Left V 12 Right V 32 Right V 13 Left 33 Left V 14 Right V 34 Right V 15 Right V 35 Right V 16 Right V 36 Right V 17 Left V 37 Left V 18 Right V 38 Right V 19 Left V 39 Left V 20 Left V 40 Left
Appendix A Experimental Phases Phase 1 Trials 1-20 Correet ehoiee (white) is 1 piece NF (0.36g) rewarded with VA piece TB (3.38g), incorrect choice (blaek) is 10 pieees NF (3.6g), experimenter assisted by pointing to correct choice and by praising correct choice Phase 2a Trials 21-25 Correct choice (white) is 1 piece NF (0.36g) rewarded with 1 pieee TB (13.5g), incorrect choice (black) is 37 pieces NF (13.32g), experimenter assisted by pointing to eorreet ehoiee and by praising correct choice Phase 2b Trials 25-30 Correct choice (white) is 1 pieee NF (0.36g) rewarded with 1 piece TB (13.5g), ineorreet choice (black) is 37 pieces NF (13.32g), experimenter assisted by pointing to eorreet choice and by praising eorreet choice Phase 3a Trials 31-35 Correct choice (white) is 1 pieee NF (0.36g) rewarded with 1 pieee TB (13.5g), ineorreet choice (blaek) is 37 pieces NF (13.32g), no experimenter assistance by pointing to correct choice or by praising correet choice Phase 3b Trials 35-40 Correet ehoiee (white) is 1 piece NF (0.36g) rewarded with 1 piece TB (13.5g), incorrect choice (black) is 37 pieces NF (13.32g), no experimenter assistanee by pointing to eorreet ehoiee or by praising correct choice Trial # Correct Correct Ineorreet Trial # Correet Correct Ineorreet Choice Choice Choice Choice Choice Choice Location Location 1 Left V 21 Right 2 Right V 22 Left V 3 Right V 23 Right V 4 Left V 24 Left V 5 Left V 25 Left V 6 Right V 26 Right V 7 Left V 27 Left V 8 Right V 28 Right V 9 Right V 29 Right V 10 Left V 30 Left V 11 Left 31 Left V 12 Right 32 Right V 13 Left V 33 Left V 14 Right V 34 Right V 15 Right V 35 Right V 16 Right V 36 Right V 17 Left V 37 Left V 18 Right V 38 Right V 19 Left V 39 Left V 20 Left V 40 Left
Appendix A Experimental Phases Phase 4 Trials 41-45 (2 weeks since previous session) Correct choice (white) is 1 piece NF (0.36g) rewarded with 1 pieee TB (13.5g), incorrect choice (black) is 37 pieces NF (13.32g), no experimenter assistance by pointing to eorreet choice or by praising correct choice Trial # Correet Correct Ineorreet Choice Choice Choice Location 41 Left V 42 Right 43 Right 44 Left 45 Left V