1) Definition of Operant Conditioning - LILT1）的操作性条件反射-定义轻快

9/10/07 VERY DETAILED Definitions in operant conditioning I. Definition of Operant Conditioning A. The Law of Effect 1. Thorndike (1911): The law of effect 2 puzzle box experiments with cats 3. first law of effect: When a response is followed by a satisfying state of affairs, that response will increase in frequency 4 also negative law of effect B Skinner: made two modifications: 1. first: problem with "satisfying state of affairs" a. reinforcement b. punishment 2. second: stated would increase PROBABILITY of response, not actual frequency C Important distinction between operant and classical conditioning 1in classical conditioning: a. CS predicts US, so react b. not HAVE to respond 2. in operant conditioning: a is a contingency in place b S+: R---> Sr or P c a stimulus may or may not predict the contingency d. in a contingency: The organism MUST make the contingent response in order to gain the consequences 3. any event which increases the probability of a response = reinforcer 4. any event which decreases the probability of a response = punisher II. Four types of consequences for producing behavior change A. Reinforcement and Punishment Defined: 1. two types of behavioral consequences: Reinforcement and punishment a using a stimulus to increase behavior = reinforcement b using a stimulus to decrease behavior = punishment 2. two types of stimuli/ways to present stimuli: a adding of a stimulus (consequence) b removing a stimulus (consequence) 3. 4 square way of looking at: Response Increase Response Decreases | | Add | positive reinforc. | positive punishment Stimulus | | Remove | negative reinforc. | negative punishment Stimulus B. Thus: 2 types of reinforcement 1. POSTIVE REINFORCEMENT a application of a stimulus to increase behavior b you RECEIVE something and your behavior increases 2. NEGATIVE REINFORCEMENT a withdrawal of a stimulus to increase behavior b something is removed or taken away and your behavior increases C. ALSO: 2 types of punishment: 1. POSITIVE punishment a application of a stimulus to decrease behavior b you RECEIVE something and your behavior decreases 2. NEGATIVE punishment a withdrawal of a stimulus to decrease behavior b you LOSE something as a result of your behavior c e.g. i Response cost ii time-out III. Parameters or Characteristics of Reinforcement and Punishment: A. Basic parameters: 1. strength of the response: a learning curve, similar to CC b monotonically increasing 2. delay of the reinforcer or punisher 3. the size, amount or quality of a reinforcer or punisher 4. extinction: a differs significantly from CC b initially is a SPONTANEOUS INCREASE in responding before get the decline in responses c also get EXTINCTION INDUCED AGGRESSION d ALSO; spontaneous recovery 5. generalization: a response may occur in highly similar situations to that of the training setting b or highly similar responses may come under stimulus control 6. discrimination: a response occurs only for specific situation in which contingency is in effect b or contingency only evokes a particular response or set of responses 7. Schedules of Reinforcement: a Schedules of reinforcment: rules that specify which response will be followed by a reinforcer b literally- a time or frequency schedule c1. Types of reinforcement schedules: i continuous reinforcement: CRF ii Partial or intermittent reinforcement: PRF B. four basic kinds of PRF: 1. fixed ratio: a cumulative recorder b fixed or set amount of responses required by the schedule c e.g. a FR 5 sets a contingency such that every 5th response is reinforced d end up with a break and run pattern of responding 2. variable ratio a vary the amount of responses required b based on some average number of responses: i e.g. like a grade point or a slot machine ii FR5: the average of every 5 responses is rewarded c ends up with steady state responding: no breaks 3. fixed interval a fixed or set amount of time required to pass before a contingent response will result in a reinforcer b can respond inbetween- but does no good c e.g. pay checks- every 2 weeks: FI 2 weeks d results in a fixed-interval scallop pattern 4. variable interval: a vary the amount of time required to pass before a contingent response will result in a reinforcer b again, can respond inbetween- but get no reinforcer c like dialing the phone when get a busy signal d results in steady state behavior 5. Characteristics of partial reinforcement schedules: a more resistant to extinction b produce more lasting behavior c avoid satiation problems C) Now can combine schedules: 1. concurrent schedules 2. multiple schedules 3. mixed schedules D) can alter criteria 1. DRL: differential reinforcement of low rates of responding 2. DRH: differential reinforcement of high rates of responding I. Introduction: A. Study of choice behavior = determining how and why organisms make choices 1. will investigate several theories/paradigms a. behavioral economics b. optimal foraging c. matching models 2. want to understand the dynamics under which organisms make choices and see if are the best choices 3. if not- where do organisms go wrong 4. biologists ignore learning factor 5. psychologists/learning theorists ignore biological/adaptation factors B. Important concept: LEARNING IS A PRODUCT OF EVOLUTION 1. in a sense, is our strongest inherited behavior 2. animals that learn can survive! 3. question for behavioral ecologists: how does behavior affect biological fitness 4. question for psychologists: how does biological fitness affect learning! II. Models of choice: A. Choice behavior and economics: Several important concepts 1. assume organism free to choose 2. choose based upon value of outcome: a. utility versus b. perceived utility 3. limitations of theory: a. not able to predict individual behavior in set situation b. making general rules, or heuristics rather than algorithms B. Understanding how and why animals make choices 1. simple choice: 2 alternatives 2. most of life is complex choice 3. how can we study choice in lab? a. T-maze: two or more choices b. Radial arm maze c. Two concurrent schedules C. Two schedules available at the same time 1. CONC 2. usually use VI schedules to obtain steady rates of behavior 3. allows us to look at choice in context 4. will read about using a COD or change-over-delay a. time out for making change b. use to prevent random responding D. ways to measure responding: 1. absolute responding: number of responses per minute 2. relative rate of responding: number of responses per minute for choice A relative to choice B 3. relative rate is a contextual rate 4. is a RATIO rate 5. allows for variation in responding across animals and settings III. Models of choice A. The matching law: 1. P1/P1+p2 = R1/R1+R2 2. this is herrnstein’s equation 3. also Baums relative rate equation a. P1/P2 = R1/R2 ab. P1/P2 = b(R1/R2) c. Yields three variables: (1) Undermatching (2) Overmatching (3) Response bias B. Undermatching 1. less than optimal performance 2. spend too much time on the poorer alternative 3. most animals do C. Overmatching 1. greater than optimal performance 2. spend too much time on better alternative 3. is not optimal: overmatching not “better”- 4. note: missing out on reinforcers 5. find when animal must pay some cost to make change in choice D. response bias 1. individual bias or preference E. e.g., left bar or right side of box IV. Molecular vs. molar theories: A. Molar theory is global 1. explains behavior over long time horizons 2. across many sessions 3. general strategy, not response by response strategy B. Molecular theory = micro 1. explains behavior across very short time horizons 2. response by response 3. explains behavior within a session 4. in way: molecular should explain molar; molar explain molecular 5. which is more important? C. Local rate versus molar rate of responding: 1. molar rate is overall responding across the session 2. local rate is the response rate on that manipulandum at that time D. basic assumption = optimality 1. assumption is that all animals optimize 2. get most bang for least buck 3. get most reward/food/whatever for least energy expenditure/behavior E. can examine matching on molar or molecular basis: 1. general overall response rate: overall GPA 2. local response rate: points in this class 3. FR/VR schedules and VI/VR schedules get different outcomes- can you figure out why? 4. ratio: rate of responding depends on how fast you respond 5. interval: rate of responding cannot change time F. BIG question: is matching a consequence of a process that maximizes reward at a molar or molecular level? a. Data support both positions b. Can molecular changes lead to molar patterns? 2. problems: Matching law not lead to an accompanying theoretical paradigm to explain WHY animals match VI. What is the primary concern of behavioral ecologists and sociobiologists? A. What is the primary concern of learning psychologists? Examine how animals learn: 1. the description and prediction of behavior preceded and followed by food 2. acquisition, extinction, etc. 3. interested in how animals acquire behaviors, maintain behaviors B. Two (behavioral ecologists and learning psychologists) have several things in common 1. Relationship between pattern of food presentation and behavior of the animal 2. Psychology: emphasizes prediction of behavior 3. Biology: emphasizes WHY behavior is important in terms of survivability C. Learning and evolution ARE related 1. Learning = instinct (as in evolution) 2. learning how to predict/control the environment is higher animals most basic instinct 3. why? D. Psychologists approach to learning behavior in animals: . 1. Studied classical and operant conditioning 2. Unfortunate consequences: a. Used relatively few species b. Limited experimental techniques/locations c. Assumed learning is highly generalized and not species specific 3. Not examined adaptive significance 4. Kamin and Yoerg argue that biology is better equipped to handle these issues- do you agree? III. Limitations on learning: A Biologists point of view! A. specifically: limitations of the matching law: 1. Matching is either very specialized or very generalized 2. not know if it is a Molar or molecular phenomenon? 3. not know if it Is it really a general phenomonen? 4. THEY Argue is NOT a general priniciple and is limited to VI schedules 5. Are they correct? Newer data say no B. Kamil and Yoerg’s example of limitations on learning: taste-aversion learning. 1. Garcia effect 2. does represent a Limitations on learning 3. One of first evidences of biological boundaries! C. foraging ecology and learning 1. Foraging ecology = how animals forage for food 2. Spatial or temporal distribution 3. Food delivery is nonrandom 4. Two important developments a. emergence of field studies of foraging: lots of settings and species b. development of optimal foraging models to explain foraging D. Optimal foraging theory as a learning model 1. N1/N2 = R1/R2 2. highly similar to matching or behavioral economics 3. Some differences: a. Search time/pursuit time b. Information about prey availabitlity c. How forager acquires information- how learned 4. Includes factors such as diet, energy output, etc. that not generally see in lab settings E. How can ecology contribute to psychology (and vice versa!): three contributions: 1. Heterosis: combination of foraging studies with learning studies 2. designing more realistic learning experiments 3. prediction of differences within species 4. prediction of differences between species 9/17/07 Herrnstein Herrnstein, 1970 Law of effect (Thorndike, 1911): behavior that is followed by a satisfying state of affairs is more likely to occur again and vice versa Note that animals improve on performance: optimize, not just repeat same behavior but make it better and more efficient Note that considers adaptation a question, and not an answer (e.g., what are adapting to; how are adapting; how know to adapt, etc) Violations of Law of Effect and “adaptation” 1. Ferster and Skinner: change in response rate between VR and VI schedules: why if just stamping in of behavior? Why do this? 2. behavior on conjunctive schedules: conjunctive schedule reinforces first response after certain number of responses AND first response after set amount of time (double criteria): CONJ FI(t) FR(n+1) a. note: for FI schedule: rapid responding is penalized b. ratio schedules: rapid responding = more reward c. CONJ schedule combines: rate of Sr is directly proportional to rate of responding only for rates of responding no larger and n/t 3. Herrrnstein and Morse (1958): a. CONJ FI 15; FR (0 to 240) b. Rate of responding decreased with increases in FR value which reduced rate of reward (Figure 2) c. Why is the animal doing this? – can’t be “stamping in” of response or wouldn’t get fluctuations d. May have to do with response strength: Reinforcement as strength: 1. making stronger link between responding and reward 2. relative frequency as measure of strength: response rate as function of reinforcer rate 3. plot proportion of responses as function of proportion of reward 4. note: is continuous measure, and not discrete trial: animal has more “choice” 5. reinforcers come on interval, not ratio schedules 6. no proportionality between number of responses and number of reinforcers- all time based 7. this becomes basis of matching law Pl/Pl +Pr = Rl/Rl + Rr Explains why this true Response strength and choice: 1. use of COD explained again 2. Shull and Pliskoff (1967): used COD and no COD and ICS as reward 3. Used findly procedure (remember what this was?) 4. matching occurred as long as used COD 5. why important: a. COD not controlling factor- response ratio was b. Extended matching to rats c. Extended matching to another Sr d. Shows several other studies that also found matching 6. Objections to matching: a. Probability models or probability learning b. Melioration models or VR schedules c. But note: here matching would be exclusive choice d. Can show mathematically can produce matching from probability learning 7. Note, however, are getting some deviations from matching under certain circumstances- need to address these circumstances from the matching law model or else matching law becomes too limited- Baum will do this next Choice as Behavior and Vice Versa: 1. Know that relative rate of responding varies with relative rate of reinforcement; must affect absolute rate of responding as well. 2. what is effect? P1 = kR1/R1 + Ro 3. makes a hyperbole function: some maximum rate of responding 4. figures on pages 255-156 5. again, provides evidence to support this 6. thus, absolute rates are affected by reinforcement rates 7. can go back and forth between two: Herrnstein equation vs generalized matching law 8. next section: explains multiple schedule and how the model fits this, we will skip that for now Behavioral contrast: make predictions about interactions: Reynolds, 1961 9/24/07 Baum, 1974 I. His generalized matching law: A. Describes basic matching law: P1/P1+P2 = R1/R1 + R2 1. Moves to: P1/P2=R1/R2 2. Notes that Staddon (1968) found can log it out to get straight lines 3. Also adds two parameters: k and a (we will use b and a) 4. Log(P1/P2) = a*log(R1/R2) + log b a 5. Or P1/P2 = b(R1/R2) 6. What is b and a? bias and undermatching B. Undermatching: 1. Fantino, Squires, Delbruck and Pterson (1972):any preference less extreme than the matching relation would predict 2. systematic deviation from the matching relation for preferences toward both alternatives, in the direction of indifference 3. what would be indifference? What value of the slope of the line? 4. slope of a of the line fitted according to equation is LESS THAN one. 5. in a sense, is a discrimination or sensitivity model: tells us how sensitive the animal is to changes in the (rate) of reward between the two alternatives 6. factors affecting undermatching: a. discriminability between the stimuli signaling the two schedules b. discrminability between the two rates of reinforcers c. component duration d. COD and COD duration e. Deprivation level f. Others? 7. Why do you think most animals undermatch? C. Bias: 1. magnitude of preference is shifted to one reinforcer when there is apparent equality between the rewards 2. unaccounted for preference 3. is experimenter’s failure to make both alternatives equal! 4. calculated using the intercept of the line 5. at least four sources of bias: a. response bias b. discrepancy between scheduled and obtained reinforcement c. qualitatively different reinforcers d. qualitatively different reinforcement schedules 6. Response bias: a. Difficulty of making response: one response key harder to push than other b. Color c. Side of box, etc. 7. difference between scheduled and obtained reinforcement: a. animal pauses, lowers obtained reinforcement even though programmed at higher rate (delivery dependent on responding!) b. thus: matching law applies only to obtained reward, if large discrepancies between obtained and scheduled, must use obtained to see animal’s preference c. some data suggests that this may not be true: animals attend to programmed or scheduled reward in social situations 8. Qualitatively different rewards: a. Matching law only takes into consideration the rate of reward b. If qualitatively different, must add this in ac. So: P1/P2 = V1/V2*(R1/R2) d. Interestingly, can get u-shaped functions rather than hyperbolas this way 9. Qualitatively different schedules: a. Use VI and VR b. Animal should show exclusive choice for VR, or minimal responding to VI c. Can control response rate, but not time d. Not “match” in typical sense, but is still optimizing D. Conclusions: 1. Matching holds up well under mathematical and data tests 2. some limitations for model 3. tells us about sensitivity to reward and bias 4. now: where would social interactions fit into this? 10/01/07 Baum and Kraft (1998) and Bell and Baum (2002) I. Baum and Kraft (1998) A. Defines optimal foraging and the equations used 1. shows how it relates to matching law (is really the same!) 2. r* = f(A,N) 3. individual capture rate is a function of amount and number of other foragers 4. r* = cA/N 5. individual rate is a function of A/N 6. so A1/N1 = A2/N2 7. rearrange: N1/N2 = A1/A2 which, in a sense, is the matching law 8. states that the ratio of the number of foragers at a given feeders = ratio of reward available at those feeders 9. is a GROUP version of the matching law a. Instead of examining matching by individiauls, look at matching for the GROUP b. So: if have 10 foragers and ratio is 4/1 should distribute 8/2, etc. B. Several key differences: 1. Key differences: is based on available and not obtained reward 2. Matching = property of organism 3. Group matching = feedback function and property of environment C. Discusses types of patches: 1. Those in which prey arrive time to time 2. Those in which abundance of prey exists over an area 3. Baum argues not important distinction: a. still have to search for prey either way; b. both search and waiting/hunting under control of consequences (prey density,) c. both are operant behaviors D. What is an operant: contingent upon consequence; 1. behavior controlled by consequences 2. Interference: social interactions 3. Should not be lower in continuous-Input patch versus standing crop 4. Still must “fight” to get prey (think of our feeders) 5. Actually may be higher in terms of face to face conflict E. What can interference effect? 1. Search and detection 2. Switching (why?) F. Tregenza (1994) suggests if low interference, then overmatching 1. Distribution of group more extreme 2. Thus greater the interference, the smaller the sensitivity to changes in distribution of resources G. How can test this? 1. vary number of foragers 2. vary size of patches II. The experiment: A. So what did they do? 1. Pigeons 2. Three kinds of patches: a. areas b. troughs c. bowls 3. travel or no travel 4. barrier or no barrier (visual block) B. Phases: 1. General foraging 2. Phase A1: separated areas; low competition with travel 3. Phase B1: troughs; moderate competition with travel 4. Phase C1: bowls, high competition with travel 5. Phase A2: adjacent areas; low competition; no travel 6. Phase A3: adjacent areas; low competition; visual barrier C. Results: 1. How analyze data: a. # pigeons in feeder 1/#pigeons in feeder 2 b. plotted against time and as function of reward D. Level of competition: (figures on page 234) 1. Areas: m = 0.79, r2 = 0.89 2. Trough: m = 0.71, r2 = .94 3. Rapid presentation: m = 0.71, r2 = 0.97 4. Bowls: m = 0.38, r2 = 0.94 5. No real differences in competition level 6. Data DO show difference in within-session variability: as session goes on, less variability, pigeons settle down, so to speak E. Phase A2: 1. adjacent areas; low competition; no travel: m = 0.61, r2= 0.93 2. Considerable undermatching F. Phase A3: Visual barrier: 1. m = 0.56, r2 =0.93; again considerable undermatching 2. Pooled these two conditions together 3. Concludes: travel improved sensitivity to reward! G.. Noncorrespondence between group results and individual performances: 1. could be aggregate of individual performances; all individuals matched, this just a mean 2. could be that some showed overmatching, some undermatching and this mean is “artificial” in terms of individual performance 3. examined this on subset of pigeons using preference proportion and switches a. found range of performances and preferences from indifference to exclusive preference (see figures 9 through 12) b. found degree of participation was consistent for each pigeon c. probability of switching or bias was inconsistent H. What does this all mean? 1. competition varied inversely with size of patches 2. as predicted, competition predicted lowered flock’s sensitivity to resource distribution 3. travel modestly increased sensitivity, as predicted 4. increase in sensitivity with travel may be due to decrease in probability of switching- switched less (why)? 5. visual barrier had no effect (why?) 6. even when competition minimized, pigeons still undermatched I. individual behavior was not necessarily correlated with group behavior- group matching is not a mirror of individual matching and group matching does not predict any individual’s behavior 1. is not emergent 2. is dynamic 3. unequal distribution for individual pigeons may reflect a. differences in individual competitiveness b. individual differences in reinforcer sensitivity c. level of participation (which also then affects competitiveness) d. pigeons interfering with one another i. switching interference ii. searching interference iii. resource sharing interference J. Now apply to what we have seen in our rats: 1. Are there individual differences? 2. Should the group match? 3. Should each rat match equally? 4. What kinds of variable should affect individual ability? 5. How will individual variability affect matching? Bell and Baum 2002 I. Assumptions of Ideal Free Distribution A. What three assumptions? 1. Perfect knowledge of patch ratio 2. Maximize based on this knowledge 3. All foragers free to enter patch on equal basis B. IDF assumes “perfect knowledge”- what does this mean? 1. know all about reinforcement situation 2. know schedule 3. at ASYMPTOTE- not learning about situation C. So: Again: why Undermatching? 1. Competition 2. Not have perfect knowledge 3. Where knowledge come from- a. past experience in relatively stable environment b. -learning rules D. Point of experiment: 1. assess Sr sensitivity to previous and current resource ratios 2. Perfect and imperfect knowledge 3. Note: used more extreme ratios than typical (made VERY salient) 4. Used more ratios (again, salience issue) II. The Experiment: A. Exp 1: Determine if foraging depends on previous experience: 1. sensitivity higher at beginning of ratio presentations in regular but not irregular sequences 2. correlations between forager ratio from final seconds of previous session to initial seconds of new session should be similar in regular 3. rapid adjustment to changing sessions- should occur in minutes B. what do? 1. 34 pigeons 2. Two foraging sites 3. Pg 183 schedules C. How analyze data: analyzed in 15 second blocks 1. Pig N/Pig N 2, Plotted sensitivity across minutes in session 3. See figure 1 on page 183: irregular low in beginning than regular D. Determine if there was Carryover from previous session 1. only an advantage if schedules are stable- 2. Found Rapid change in irregular groups 3. How much carryover was examined: a. Figure 2: correlation between sessions 1. high for regular 2. Low for irregular 4. Figure 3: shows sensitivity lower for irregular (pg 186) E. Discussion: 1. Flock adjusted to changing resources 2. Hypothesis were found to be true III. Experiment 2: A. What did they do? 1. Isolated indoor coop 2. Can’t see food delivery 3. Same basic setup 4. Basically same experiment but automated B. Results: 1. Basically same results 2. More stable results C. Discussion 1, Even when unstable schedules, still match via distribution of pigeons 2. group adapts at different rates depending on “information” 3. Do use past experience IF it provides knowledge D. Conclude that this is adaptive- what mean by that? 10/08/07 I. Madden, Peden, and Yamaguchi (2002) A. Introduction to study is all about optimal foraging 1. reviews the premises 2. N1/N2 = A1/A2 3. habitat matching B. Because foragers in resource site must share resources available in that site 1. deviations from matching mean individuals in 1 group obtain more resources than another 2. e.g.: 50 units of food available a. 40 animals in site 1 b. 60 in site 2 c. Site 1 animals: 1.25 units/animal d. Site 2 animals: 0.83 units/animal 3. model states that therefore some animals from site 2 move to site 1 to even out and equalize ratios C. notes that this is simply the matching law 1. ideal matching 2. bias 3. under and over matching 4. reviews Baum and Kraft 5. NOW: do humans do this? D. Human discrete trial procedure 1. group of humans 2. choose 1 of 2 resource sites 3. show red or green cards 4. can move freely 5. found matching 6. problem: real life foraging NOT a discrete trial a. so: purpose of experiment is to examine humans b. but in free foraging situation E. What do: 1. experiment 1: a. 12 college students b. Could earn $30 and $10 for earning points c. Seated in desks in a circle i. Two cards: red and blue ii. Score sheets iii. Earn as many point as possible: a. Allocated different amounts of points to each card b. But have to share them among people who show same card c. So: if red = 100 and 10 people choose red: only get 10 points d. Varied ratio of red to blue points iv. At “choose now”: showed card, could keep changing a. When no switching for 5 secs, recorded answer b. Counted red/blue cards, divided and recorded points c. Put cards face down again, and did again d. About 20 trials per condition d. results: i. good interobserver agreement (.98 to 1.0) ii. compared predicted and obtained “matching” ratios a. intial decision b. final decision iii. initial selection: a = .4, r2 = 0.55 iv. final selection: a = .92, r2 = .99 v. no bias e. initial selection much different than final selection- they balanced themselves out to get most points! i. Why? ii. More knowledge of others actions: counted and observed others iii. Shows this foraging is fluid- as predicted 2. experiment 2 a. no talking or switching! i. Same set up ii. Only initial card selection counted iii. Continued trials in a condition until selection was stable (stability criterion) b. Results: i. Good IOA: 0.97 to 1.0 ii. Now found better matching: a = 0.82, r2 = 0.98 c. Why? i. Initial selection now more important ii. No more socialization to determine choice iii. Chose according to odds, improved with trials iv. Responding to ratio of reinforcement, not social cues 3. Experiment 3: a. Now used conc VI VI schedules rather than discrete trials b. Tables at each end of classroom where go to earn points i. Red paper zone ii. Blue paper zone iii. Must move to enter a zone (note- no chairs in room, subjects did sit on table or near a zone, though!) iv. Observers recorded placement of participants time in “reinforcer zones” v. Had to be in a zone when points awarded according to VI schedule c. Results: i. IOA results good again: 0.98-0.99 ii. Movement of participants tended to stabilize after initial movement period of about 20 minutes iii. A = 0.71, r2 = 0.99 iv. Not as good as cards, but not bad matching d. Group was sensitive to changes in conditions i. Less sensitive to reinforcement- why? (was harder) ii. Interesting that initial movement, then ratios of individuals in each zone settled down iii. Also: if more sensitive to reinforcer magnitude (absolute amount of points) rather than ratio, follow dispersement of points, even when at lean side F. Conclusions: 1. humans match! 2. sensitive to group dynamics 3. Baum and Kraft found a. Individuals do not match time or response allocation to distribution of reinforcers obtained by group b. No regular patterns of switching across sessions c. No regularities or preferences for one resource site over another across sessions or conditions d. No tendency for human participants to be consistently higher point earners across sessions e. BUT: looked at MOLAR level 4. Madden, et al: a. Looked at more molecular level b. Did find shifts in switching WITHIN a session c. Showed melioration: moving to the better source d. Suggests are both molar and molecular changes that lead to matching! II. Reed, Critchfield and Martens (2006): Football and matching A. Again, note that matching law can be broadly applied 1. widely used in animal research 2. optimal foraging 3. human research a. employee absenteeism b. teen pregnancy c. classroom behavior 4. want to know if it affects sports behavior (why? Because it is there!) a. sites several studies b. Vollmer and Bourret (2000): i. Matching and basketball ii. Choosing 2 point vs 3 point shots iii. Matching law described decision making for taking a shot iv. Bias towards making 3pt shots (why?) v. a = about 1.0 5. why choose football? a. Play calling = individual behavior i. Quarterback ii. Offensive coordinator and head coach iii. Highly skilled b. When calling play, consider success/failure of previous attempt in decision for next play c. Individual differences in play-calling patterns (throwing vs passing teams) d. Focus at team level 6. general method a. data obtained from NFL b. primary data: i. number of passing/rushing plays ii. net yards gained c. several characteristics i. plays categorized as rushing or passing based on what occurred rather than what was called (no way of knowing that) ii. sacks = failed rush play iii. yards gained = completion even if fumble after catch d. fit data to matching equation i. ratio of yards gained through passing vs rushing used as predictor of ratio of pass plays/rush plays called 7. season aggregate league outcome a. a = 0.725, r2 = 75.7; b= -0.129 (favor of rushing) b. historical comparisons: i. 1975-2005 ii. 2004 fell out of typical range iii. R2 decreases about 4%/year across years, suggesting more variability in play calling iv. Why? a. Shift in rules designed to favor passing b. Free-agency rules c. Salary caps c. comparison with other leagues: i. NFL Europe: 0.619. 82.1 ii. CFL: a = .544, r2 = .567 iii. Arena Football: a =.56, r2=.784 iv. United Indoor Football League: a = 61.3, r2 59.8 v. National Women’s football association: a = .55, rw = .709 vi. NCAA Atlantic Coast: a = 0.63, r2=.809 vii. NCAA Western: a = .868, r2=.946 viii. NCAA Mid-america: a = 0.509, r2= ix. 634 x. R2 = .57-.95 xi. Generally good fits xii. 6 of 9 leagues: favored passing rather than rushing xiii. CLF: rushing rather than passing (turnover risk?) d. conditional play calling: i. examined specific circumstances ii. examined down number (1,2,3) iii. how does matching change? a. a = decreasing with down b. less likely to pass with increased down c. is this surprising? 8. Game by Game outcomes: a. Regular season games b. Preseason fits relatively poor: a = .43 c. Later in season: better fits: a = .58 d. Post season slightly better: a = .59 e. Why? 9. Game by game individual team outcomes a. Does matching fit predict team success? b. Teams show some variability in fit to equation (see pg. 290) c. Interesting: no relationship with lay descriptions of teams i. Colts known as passing team, but showed more rushing bias than Atlanta, which was a “rushing team” ii. Has to do with ratio of called plays to success iii. Bias = avoidance of risk of turnover iv. Many influences on play calling a. Coach vs. quarterback b. Team structure may differ d. Better (more successful) teams showed steeper matching functions i. If matched better, were better team ii. Suggests that matching impacts winning! iii. This was significant! R = -.579, p=.0007 a. Why negative? b. R between matching and LOSSES 10. cause and effect questions: a. play calling influences yardage gained b. but yardage gained influences play calling! c. Real world influences d. Many unaccounted for factors e. What is contingency here? i. Matching law assumes a contingency ii. What is the contingency for players/coaches, etc? f. Data suggest that better sensitivity to reward = better success g. How else can use matching law in real world? 10/15/07 Farmer-Dougan and Dougan contrast!Behavioral Contrast in group foraging I. Defining behavioral contrast: A. Many kinds of contrast 1. simultaneous 2. successive 3. behavioral 4. consummatory B. multiple schedule contrast 1. schedules flip flop 2. classic contextual phenomenon 3. predicted by matching law (although not explained) 4. also explained by ideal free distribution model C. explains difference between 2 1. matching law = individual organism 2. IDF = groups 3. little research has compared models D. Defining contrast 1. Reynolds (1961) 2. inverse relation between rate of responding in one component of multiple schedule and rate of reinforcement provided by the other component 3. three schedule series a. baseline b. contrast c. baseline recovery 4. positive contrast: a. increase in response rate for constant component relative during contrast period relative to rate of response in component in baseline phases b. responding increases as other, changed component decreases 5. negative contrast: a. decrease in response rate for constant component relative during contrast period relative to rate of response in component in baseline phases b. responding decreases as other, changed component increases 6. use multiple schedule equation to estimate: a. P1=(kR1)/(R1+mR2+Ro) b. N1=(kA1)/(A1+mA2+Ao) 7. two experiments examined multiple schedule contrast using groups II. Experiment 1 1. Method a. 5 rats b. open field c. three schedules i. baseline : VT 15 VT 15 ii. contrast: VT 15 VT 900 iii. recover: VT 15 VT 15 d. measured: i. number of 15 sec intervals in each feeder for each rat ii. number of pellets consumed by each rat iii. used time sampling B. Results: 1. Figure 1: plots number of intervals in feeder as function of three schedules 2. good positive contrast for all 5 rats (confirmed by stats) 3. figure 2: proportion of rats in feeder a. some contrast b. contrast like effect 4. BUT: look at Figure 3 a. Pellets consumed DID NOT Follow the contrast pattern b. Note: had to decrease in middle condition c. Appears were following SCHEDULE rather than number of pellets consumed d. Very intriguing: doing what was scheduled, not what received! III. Experiment 2: Negative Contrast A. Method 1. 5 rats 2. open field 3. three schedules a. baseline : VT 60 VT 60 b. contrast: VT 60 VT 15 c. recover: VT 60 VT 60 4. measured: a. number of 15 sec intervals in each feeder for each rat b. number of pellets consumed by each rat c. used time sampling B. What happended? 1. Figure 3 2. some negative contrast for all 5 rats (VERY hard to get) 3. not significant, though 4. group: negative contrast IV. Overall A. Group vs. individual effects 1. individuals showed contrast 2. group showed contrast 3. followed schedule, not what consumed B. shows group foraging approach worked! 1. fit models from literature 2. looks like data from individual animals in literature C. most interesting: 1. lack of correspondence between pellet ingestion and distribution of behavior 2. gives idea of what rats “attending” to a. not base behavior on what happened b. based behavior on longer time horizon- the schedule c. intriguing because shows i. awareness of schedule constraints ii. ability to change behavior based on future/predicted outcomes D. differences in this experiment 1. type of schedule: VT vs VI 2. implied operant 3. groups 10/22/07 Rationality and one feeder vs. two feeder defense strategies Study Guide for Readings for Week #9: Are animals rational? Humphries, S., Ruxton, G.D., & Metcalfe, N.B. (1999). Patch choice and risk: Relative competitive ability is context dependent. Animal Behaviour, 58, 1131-1138. I. What is relative competitive ability? How is it important for ideal free distribution/optimal foraging theory? A. Relative competitive ability = ability of an individual to compete for resources in relation to abilities of others within a group 1. Important for understanding dominance and resulting hierarchies 2. Also important test of ideal free distribution B. IFD predicts 1. animals of equal competitive ability who are free to travel within/across patches w/o cost 2. and have absolute knowledge 3. should exploit patches where they are able to maximize resource gain C. If have more than one resource source: 1. result should be density dependent effects which lead to equilibrium of distribution 2. such that all patches proved equal resource acquisition rate 3. and is little movement across patches (better to stay) D. How test? 1. Can vary type of patch to see differences in competitive status 2. determine how this affects IFD, but little empirical data- this is what they did II. What was the purpose of their study? A. To understand 1.relative competitive ability of individuals in free choice between two feeding situations 2. used a variance in food input rate about a set means as differentiating factor between two patches; 3. examined data with single patch and two patch situations B. Thus: they 1. Studied relative foraging ability in each of two foraging situations 2. first a one feeder situaiton 3. then compared to data with two-feeder situation C. Describe their method: 1. used tilapia trout as subjects 2. Tanks with two foraging situations: 3. used “trout pellets” as reinforcers! 3. Single patch feeding: a. constant feeder (FI schedule) b. variable feeder (VI schedule) 4. Compared to two-feeder situation: FI and VI III. What were the results? A. ranking in the single patch vs. ranking in the two-patch condition? 1. Rank order in VI single patch vs FI single patch did not change 2. High correlation between rank across two schedule situations 3. Was significant difference in food intake: more intake at variable condition, but was group and not individual effect 4. NO correlation between rank in individual setting and two-feeder setting! B. For two feeder situation: 1. Was clear ranking, but not same as in one feeder situation 2. Again, difference in food intake across two feeder patches 3. but more in CONSTANT feeder this time 4. Significant correlation between time spent in patch and intake: 5. more time = greater intake C. Relation between ranking in one patch and two-patch conditions 1. ranking in one patch of the two-patch condition DID predict rank in the other patch? 2. There NOT not a significant correlation between rank and size of the fish 3. no significant correlation between weight and which patch they favored D. Did the fish optimize or match as a group? 1. Yes! After initial settling down period (why?) 2. Again, more fish in constant than variable 3. But: no difference in individual intake in two patches E. Did switching rates correlate with rank? 1. YES! More switches correlated with lower rank! 2. Switching in two feeder not related to switch outs from one feeder situation! 3. Negative correlation between switches from center area to one of patches F. Did relative competitive ability remain constant in the one patch or two patch condition? 1. seemed more constant in one feeder 2. in two feeder: more variability 3. why? IV. Risk aversion and patch foraging A. Several interesting phenomena occurred 1. Within single patch: dominant fish able to defend feeder 2. In two feeder situation: not able to defend both at once 3. variability of reward, dominants may not have risked running off less dominant fish in that might miss a reward 4. Best individuals were not able to monopolize feeder B. Why lack of correlation between single and double feeder? 1. Reduced competitor density at two patches 2. Changes in relative competitive ability – constant vs variable may result in different relative competitive ability across animals 3. C. Risk aversion to variable patch: 1. cost of switching patches 2. stay at constant rather than risk missing reward in variable patch 3. Changes/improvements across time in relative competitive ability (doubtful given stability in single feeder situation) Schuck-Paim, C. & Kacelnik, A. (2002). Rationality in risk-sensitive foraging choices in starlings. Animal Behaviour, 64, 869-879. What is behavioural coherence? Why is it important? Subject’s preference system should behave rationally Preferences are consistent across contexts, regardless of what these preferences are or the processes by which the subjects make decisions Completeness: well defined preferences between all possible pairs of options Transivitiy: if A > B and B>C then A>C Regulatity: increase in number of choices never increases choice probability of given option Why is it important to study breaches of rationality? Few normative frameworks in biology that are compatible with violations or breaches Need to know how prevalent Understand conditions under which occur What are two possible accounts for violations of rationality? Assume rationality; question: is that really the case? Subjects unlikely to attend to all alternatives in any choice set; as attending changes, in a sense the choices change (really is a “mistake”) Organisms unlikely to make decisions by assignments of worth in a one- dimensional scale, but use multidimensional scale Adaptive value of heuristics; but heuristics are not algorithms! What generally works doesn’t always work! Situations are complex Define background context and choice context. Depends on the setting: what is framework of choices? Matching law really emphasizes this!!!! Background context = effect of set options known to subjects, whether or not they are present at time of choice Choice context = effect of number of options at moment of choice What was the purpose of this study? investigate the extent to which inclusion of third alternative into binary choice affected starlings’ preferences between original foraging options explore generality of previous research see if animals behave as predicted What were the authors’ hypotheses? Three settings: binary choice; addition of third choice; two choices with ghost third choice (randomly selected two of three choices to present) Three variance levels: none, medium, high Describe their method: 18 wild caught starlings Experimental pigeon cages, essentially (With perch) Food reward varied in time presentation/delay of reward Trained with choice or no choice procedures (really forced choice) What was the procedure? Experiment 1: training was binary Testing was binary NM MH NH groups Experiment 2: training: trinary Testing: trinary Experiment 3: training: trinary Testing: binary Calculated percentage choice: For binary: 50% is equal For trianary about 1/3 Describe context dependence (pg. 873) in words. Give an example. Condition asserts that the proportion of choices for an option cannot be increased when a new option is included in the choice set Possibility of violation if proportion of choices allocated for new alternative (z) is large. This leaves little responses for x and y, which means must fight to maintain ratio of choice between x and y example: Make 100 responses: 60 to x 40 to y Now add Z Make 100 responses 40 to Z that leaves 60: must make 36 to X and 24 to y Make 100 responses 90 to Z that leaves 6 to X and 4 to y Things to think about: Describe independence from irrelevant alternatives (p 873). Can you think of an example? Choice does not consider irrelevant alternatives Describe the assumption of scalability or proportionality. Can you put it into your own words? Is preference scalable or orderable: is this order really able to be scaled (What measurement system?) How did the authors’ analyze their results? What did they “test”? Could the starlings detect correctly the timing of delays associated with the High treatment? YES peck rates did not really vary at time bins when food was potentially due for medium treatment peck rates for the None condition showed increasing rate across time Can conclude about the starlings ability to recall the time occurrence of rewards: Able to recall, perception of each option not distorted by number of foraging alternatives experience or available. None to medium least extreme medium to high most extreme differences Starlings DID prefer variability in the delay of reward Did show context dependence and some Transivity showed strong preference for variability in delay of reward when choice set included high variance option (remember risk avoidance) More moderate difference for none and medium Did show regulatory though Did the authors find violations of rationality? Yes and no depending on clarity of situation, starlings were rational if situation unclear or used extremes, was less rational Things to think about: Why did I have you read these articles? Are optimal foraging and matching similar? How to these studies help us understand the choices made by our rats in our foraging box? 10/29/07 Arcis and Desor and Aparicio Arcis, V., & Desor, D. (2003). Influence of environment structure and food availability on the foraging behaviour of the laboratory rat. Behavioral Processes, 60, 191-198. I. Arcis and Desor: environment vs. food I. What are the two important questions for foraging strategy? A. Diet choice B. Which prey item to consume C. Patch exploitation: When to stay or leave a patch II. What is the advantage for an individual animal in gaining information about predation? A. Optimize energy intake 1. Should exploit preferentially areas with abundance of resources 2. But other factors: a. Thermoregulatory costs b. Reproductive activities c. Risk sensitivity d. Presence of competitors e. Predation risk f. B. Individuals able to use info about predation risk gain selective advantage 1. Environmental cues 2. Direct and indirect C. What three experimental situations were used in the experiment? 1. Changes in physical structure (experiment 1) 2. Food availability (experiment 2) 3. Both (experiment 3) D. Explain the methods: 1. Subjects: wistar female rats 2. Apparatus: maze using 140 bricks 3. Procedure: used food reward: chocolate flavored puffed up rice 4. Mixed up food density vs. brick density E. Describe each of the three experiments; how were they the same and different? 1. Experiment 1: a. Control session first: Identical food/brick densitities b. Then second patch presentation c. Varied density of bricks i. 1/3 ratio ii. Closed patch vs. open patch iii. Same food density 2. Experiment 2 a. Control session again b. Variation of food density c. Again 3:1……81 versus 27 items d. Brick density identical 3. Experiment 3 a. Two food parameters b. Two brick densities c. See which they preferred F. Behavioral observations 1. Latency to enter maze after run onset 2. Time spent outside maze 3. Time spent in maze 4. Distance covered 5. Number of food items eaten III. What were the results for the: A. Control 1. Almost all time spent in maze 2. No differences observed between 2 mazes or patches for any measure B. Experiment 1 1. Spent less overall time in maze 2. More time in “closed” maze 3. Ate more in closed maze 4. Moved more distance in closed maze C. Experiment 2 1. Least amount of time in maze 2. More time in higher food density 3. Covered more distance in higher food densitity 4. Ate 3 times more in high food densitity D. Experiment 3 1. Similar amount of time as exper 2 2. Spent more time in closed patch 3. Ate similar numbers of food items in 2 patches 4. Did not differ in distance covered E. What were the authors’ conclusions. That is, why was their experiment important for foraging theory? 1. Exper 1: avoided open patch 2. Exper 2: ate at richest patch 3. Exper 3: trade off 4. Compared results of 3 to 1 a. percentage of time in open patch higher when offered greater food b. moved more in open patch when offered greater food c. ate similar amounts , though d. ate less in closed part when offered less food than open than compared to experiment 1 F. Conclusions: 1. environmental structure played role in foraging condition 2. suggested was clue for predation risk 3. again, patch food density was important: preferred more food 4. when given combo: trade-off a. looks like safety has slight advantage b. would this change if more hungry rat? If stronger predation clues? III. Apricio, C.F. (2001). Overmatching in rats: The barrier choice paradigm. Journal of the Experimental Analysis of Behavior, 75, 93-106. A. What is the barrier choice paradigm? What does this paradigm have to do with matching/foraging? 1. Locomotion effects on choice when must travel between choices 2. More costly travel = overmatching B. What was the purpose of this paper? What were the hypotheses? 1. Two phases: different barrier heights 2. Phase 1: shaped to bar press THEN 30.5 cm height barrier 3. Phase 2: 45.7 cm barrier added (15.2 cm higher) 4. Hypothesis: higher barrier = greater overmatching 5. Looked at time and responses C. What were the results? 1. Phase 1: Responses time allocation Rat 53: 1.36 0.10 1.15 0.07 Rat 59: 1.11 -.20 1.24 -.28 Rat 60: 1.30 -.13 1.04 -.03 Rat 61: 0.88 0.09 0.87 0.07 1. Phase 2: 30.5 cm Responses Time allocation Rat 51 0.96 0.10 0.97 -.03 Rat 52 0.99 0.14 0.99 0.01 Rat 54 0.95 -.25 0.89 -.17 Rat 55 0.93 -.16 1.08 -.14 Rat 56 1.02 -.02 1.01 0.09 Rat 57 1.06 -.07 1.06 -.01 Rat 58 0.82 00 1.01 -.10 Rat 62 0.79 0.28 0.86 -.01 4. Phase 3: 45.7 cm responses Responses Time allocation Rat 51 1.14 -.13 .93 -.02 Rat 52 1.28 0.12 1.03 0.15 Rat 54 1.155 -.05 1.05 0.01 Rat 55 1.11 -.02 1.08 -.21 Rat 56 1.29 -.01 1.21 -.01 Rat 57 1.22 0.14 1.07 -.03 Rat 58 1.05 0.01 0.97 .07 Rat 62 1.118 -.23 1.05 .04 D. What is the role of locomotion (that’s moving and traveling) for matching/foraging? 1. 45.7 cm barrier: ratios of residence times overmatched ratios of obtained reinforcers 2. Raising barrier height increased degree of overmatching 2. What do these results mean for foraging and matching? Choice situations with costly locomotion lead to overmatching 11/5/07 Social factors in foraging! I. Social Foraging A. Capuchins in the wild 1. capuchins = new world monkeys a. Argentina national forest b. Form large social groups- extended family 2. foraging changes based on a. size of the family b. spatial layout of the food source B. several concepts in foraging studies that we haven’t talked about 1. higher animals (not rats, really) 2. finder’s share: a. share that the finder gets b. Kleptoparasitism: i. Stealing from others in your troop rather than foraging for yourself ii. Mooch 3. stochastic models for information sharing of food resources a. produce scrounger model in game theory b. animals can use 2 tactics: i. producers: actively search for food and find food ii. scroungers: parasitically exploit food from producers (mooch) 4. number of producers and scroungers varies by food delivery rate: a. larger payoff for scrounging than producing: increase in scroungers b. larger payoff for producing: increase in producers c. finder’s share and what the finder does with the food it eats: i. consume a lot, then share ii. consume little, share immediately 5. who the finder is depends on social status in the troop/family C. experiment: 1. examine three things: a. how does rate of food affect finder’s share, producers and scroungers b. how does social status play a role in scrounging (kleptoparasitism) c. what social factors as well as food factors affect finder’s share 2. go to the forest: a. capuchin troop- radio collared as well as observed b. observe in 5 min bins i. spatial location of each monkey ii. who is eating iii. who was there first iv. who steals from whom D. what did they find? 1. positions a. Alpha Male: smack in the middle- anterior central position b. Alpha females: periphery of alpha male c. Other females and children were behind alpha male d. Larger male juveniles are out in front and periphery- scouts e. Send scouts out to look for food 2. juvenile male arrives at food platform a. “finder”: eat first, then call others b. When alpha animals arrived, steal the food from the scouts c. Babies and moms were allowed to eat after the alpha male, but before the other scouts 3. several factors that affected how soon the scout called out: a. richness of the food patches: more food, called sooner b. status of the scout: higher status, ate less of finder’s share, called sooner c. punishment dealt out if caught with finder’s share 4. this also occurs with sexual behavior a. genetic diversity issues with primates and monkeys b. if alpha male only mates with females in troop= low genetic diversity c. alpha male will kill any offspring that he thinks are unrelated to him d. with genetic testing: turns out not all the kids are his! i. Alpha female has his offspring ii. Beta and lower females do NOT e. alpha female distracts the alpha male while the beta females run off for fun in the woods! 5. take home message: a. animals optimize! b. Get the most food in any way they can, even if it means cheating! (and for sex too!) II. River Otters A. Why form a social group? 1. protection from predators 2. find food- cooperate you might increase intake B. Mustelids: 1. aquatic carnivores 2. otters, seals, seal lions, orcas, dolphins C. two types of prey available for river otters: 1. seasonal 2. pelagic fishes: schooling fishes (groups of fish)-salmon! 3. intertidal demersal organisms: tide pool crustaceans (crabs, etc.) 4. how catch: a. pelagic fish: group behavior b. crustaceans: alone 5. gender differences a. males are bigger, stronger, better swimmers b. females: more parenting, smaller, and have smaller home range c. morphology is different (their body structure and physical behavior) B. hypotheses 1. seasonal changes in social groupings 2. differences by gender 3. social group should drive feeding success D. method: 1. Prince William Sound in Alaska 2. watch: river otters a. radio collared b. catch some of them E. results: 1. group size changed by season: a. more social grouping when pelagic fish available but for MALES ONLY! b. 48% of females remained solitary, only 24% of males did 2. males showed more variety in diet than females a. ate more pelagic fish during fish season b. gave them an advantage- pelagic fish are better food source c. males had better overall nutrition d. social isolate males were weakest, lightest and least able 3. if anything for females, social grouping varied with offspring a. more alone if had babies b. more social when no babies F. take home message: 1. food getting depends on size, gender 2. varies with season 3. strategies may vary with season, but interact with gender G. it all boils down to energy expenditure vs. energy recover 1. psychologists don’t deal with this 2. biologists do 3. how much effort do I put out for how much gain 4. optimize: get most gain for least output 11/12/07 Species Differences I. Bates and Chapell’s guppy study! A. social transmission 1. psych term: social learning or modeling 2. do animals communicate via modeling 3. can animals learn via modeling 4. test: a. model models a behavior b. observer watches c. test to see if the observer does the new behavior B. guppy behavior 1. shoaling: hugging the sides of the shoreline to avoid predators 2. shoaling may or may not be the most optimal path to reinforcer 3. technical name for the social learning: allomemesis C. guppies: 1. 12 founders 2. 8 experimental group 3. 8 control group 4. 7 spares 5. tanks: a. divided in half lengthwise b. PVC pipe with a trap door c. Blue end vs red end d. Short route and a long route (think feeder 1 to feeder 2) 6. reinforcers: freeze dried blood worms D. train the founder group: 1. train for short or long route 2. experimental group: watch the founder group 3. control group: just swim 4. test: which way do the observers swim? E. couple of side questions: 1. no actual predator threat 2. optimal route: short route 3. safer route: long route 4. which way should control group swim?: depends F. results: 1. control fish: liked the short route 2. observers: a. followed whoever they observed, generally b. trade off: i. typically, they liked the shorter route ii. if observed longer route, more likely (but not perfect) to swim the long route G. Why important for us? 1. social interactions AND social learning 2. maybe our rats learn and observe from eachother 3. if one rat runs to another feeder: the other rats ……..follow sometimes 4. could social learning affect competition which would then affect matching? II. Furyama: Hamsters and matching A. why hamsters? 1. not rats! 2. unique feeding pattern: pouch their food rather than immediately ingest 3. hoard food 4. do hoarders match? Are they sensitive to changes in the reward ratio if they have this stockpile? B. did a matching with hamsters 1. 3 hamsters: 1 girl, 2 boys 2. operant boxes: lever press 3. 80% deprivation 4. several conc VI VI schedules with COD C. results: 1. hamsters match! a. 0.87, 0.96, 0.82, M=0.87 b. Little bias: 0.02, 0.22 and 0.05 c. Great fits: 0.99, 0.87, 0.82 2. so?!?....tells us that hamsters were sensitive to payoff differences a. pouching does not replace “collecting” b. optimize the amount gained in the pouch III. Gulley: species differences A. examined SD and LE rats 1. large set of data that show that strain differences may affect cocaine and drug of abuse sensitivity 2. no one really examines baseline behavior 3. may be a model for human differences B. different behavior categories to classify rats: 1. low responder or high responder to cocaine (or any DA agonist) 2. we know that LCR are qualitatively and quantitatively different than HCR a. show lower basal behavior b. show differences in physiological response to cocaine,etc. C. two kinds of tasks 1. open field behavior including locomotion and rearing 2. operant box: progressive ratio schedule and look for the break point 3. two kinds of rats: LE and SD 4. two kinds of responders: HCR and LCR D. what happened: 1. open field first: a. strain: LE different than SD i. yes ii. no difference in locomotion iii. significant differences in rearing iv. also differences in LCR vs HCR for rearing b. for rearing: LE are always more active, but the HCR are most active 2. operant PR schedules: a. SD rats took longer to learn (light as cue) b. High responders took slightly longer to learn c. SD rats took longer to reach a break point than LE rats E. how could this apply to humans? 1. know that LE have different reactions physiologically to cocaine and other drugs of abuse than SD 2. vision differences 3. body type differences 4. could we have different kinds of humans? a. Type A vs Type B b. Type T personality: i. Thrill seeker ii. More likely to become addicted to DA activity