Define terms where needed (assume your reader is naïve and has zero background knowledge). Aims: Experiment 2A, Stimulus Control–Can we place a behavior under the control of a discriminative stimulus? Experiment 2B, Chaining–Can the discriminative stimulus (light) itself have reinforcing properties? Can the discriminative stimulus reinforce a more complex behavioral chain (adding on the left lever)? Methods (5pts):Describe how data was collected (refer to modified APA handout) Subjects: subject information (n, sex, species), housing, water and food Materials: Skinner box and reward information Procedures: Your handouts should provide you with enough information Experiment 2A: Discriminative control – include information about the design and your specific parameters. Include information regarding acquisition, extinction, and re-acquisition phases Experiment 2B, Chaining–include information about the design and your specific parameters. Not the number of days we trained the animal (March 12, 19, 26, & April 2nd). Be sure to mention that the rats had previously been trained to lever-press, either in the procedures section or in the subjects section Results (15pts):APA reporting of your results, you must include graphs (at least 2, one for each experiment). You must include each of the phases below. If it is easier to report in seconds rather than minutes feel free to do so. Experiment 2A: Discrimination: report the number of sessions and compare the time to acquire vs. extinguish, time to extinguish vs. reacquire and time to acquire vs. reacquire; to analyze learning of the discriminative stimulus, either report the average latency or response rate (in SDor in S∆) for each session of acquisition or calculate a discrimination index Experiment 2B Chaining: describe your results and compare to what we have done with the rats if possible; Include the results from both rat 1 and rat 2 You only need to present the results from the first (March 12) and last (April 2nd) and extinction (May 19th, only rat 1) chaining periods. This will give you enough information to see if the animal had learned the chaining process. Questions to answer in these results section–Did the rate of lever pressing increase between the first and last chanining periods for both rats? Did the duration of the inter-trial interval (ITI) go down for both rats between the first and last chanining sessions? Did the rate of left lever pressing go down during extinction? did the rate of left lever pressing go up during reacquisition right after? Note: We collected a LOT of data for the training session. Much more than can be expected to be used for the purposes of this class. Thus for this study what we are going to do is 1) Compare the overall rate of left and right lever presses during the trial and ITI on the first session March 12, with the last session, on April 2 for both rats and B) Compare the rate of left and right lever presses during the trial and ITI during before, during extinction and during acquisition for May 19th for rat 1. In particular in your report I want you to make a table or bar graph and interpret the differences in the discrimination index, rate of left and right lever responses during ITI & Trial for A) both Rat 1 and Rat 2 during May 12th and April 2nd and another table for just Rat 1 before, during extinction and during acquisition during the May 19. If you don′t understand how to do this please contact me and schedule a time to meet during my office hours. Graphs: You need at least one from each experiment (＝ minimum of 2). If you choose multiple within an experiment, you may just want to show one from the first session and one from the last session if you completed more than two sessions to show the difference in behavior from beginning to end of training prior to extinction. You can combine multiple graphs on the same page and refer to them as one figure. However, you should label each cumulative record in the combined figure separately, such as (a) (b) (c) or use desсrіptive labels depending on what you want to show. Remember to include the scale on the figures showing cumulative records. (see example figures at the end of this document) Types of figures: Cumulative records, plots of individual latencies, frequency distributions of latencies It is up to you what sort of figures and how many you put in your paper (your paper should at least have two figures, one from duration and one from discriminative control) For all figures, it is important to remember that you must describe in the Result section what the data in a given Figure show. What are the trends, what happens as a function of training, what happens as a function of training, what happens when you change an independent variable (i.e., what happens when you increase duration criterion, what happens when you introduce a discriminative stimulus, what happens when you introduce ITIs), how quickly does the behavior recover after extinction, etc. You must refer to the figures in your text, point to specific numbers, and point to trends in the data but withoutrepeating all or nearly all the numbers on the figures in the text. Guide the reader through the results; the main idea is to use the actual results to support what you write. Discussion (10pts):Include a summary of your study (including aims) and findings, compare with previous findings from your library search that were discussed in introduction (can also include additional research), address limitations, and discuss what the results mean and implications from research like this (refer to modified APA handout for how to handle 2 experiments) References (.5pts) TOTAL ＝50 points There will be point deductions if the minimum page limit is not achieved (it is okay if you go over 12 pages). How to Analyze Results from Experiment 2 Data Analysis: There are several ways in which you can analyze the results from this experiment 1. You should calculate the average latency for each session of acquisition of discrimination and plot the average as a function of sessions. The average latency can be obtained by averaging all the printed individual latencies in the raw data files (not including trials where no response occurred). You can also calculate the average latency by dividing the total time in SD by the number of trials. However, note that if you have trials without a response, you must subtract that time from the time spent in SD because if the rat does not press, it is not a latency to SD. For example, if you have 20 trials but the rat didn’t press on 2 of them (maybe in the first session) and your total SD duration is 300s then you have two times 60 s in the SD duration that need to be taken out (the light stays on for 60 s if the rat doesn’t press in a trial). The calculation therefore is (300 – 120)/18 ＝ 10 s. You divide by 18 and not 20 because the rat didn’t press on two of the 20 trials. So the average latency is 10 s. You will most likely see that your latency is a decreasing function of the number of sessions of training and that the latency will stabilize at some value (probably between 2-5 s). 2. You should calculate the response rate in SD (trial periods). Simply divide the number of responses in SD by the cumulated time spent in SD. Remember that the time unit is 1 s. To express the response rate as responses per minute you should multiply by 60. For example, if your responses is 20 and the time is 80 s, your response rate is (20/80)*60 ＝ 15.0 resp/min. You should plot your response rate in SD as a function of sessions, as you did for the latencies. You will most likely find that the response rate in SD is an increasing function of the number of sessions of training. Note that response rate is expressed as responses over the period of opportunity to make a press. Thus, the response rate calculation includes SD periods where the rat did not press the lever. 3. You should calculate the response rate in SΔ (ITI periods). Simply divide the number of responses in SΔ by the cumulated time spent in SΔ. Remember that the time unit is 1 s. To express the response rate as responses per minute you should multiply by 60. Thus, if your SΔ responses is 120 and the SΔ time is 840 s, your response rate is (120/840)*60 ＝ 8.6 resp/min. You should plot your response rate in SΔ as a function of sessions, as you did for the latencies. You will most likely find that the response rate in SΔ is a decreasing function of the number of sessions of training. You can plot SD and SΔ response rate as two lines on the same graph, then they are easy to compare. 4. A common way to analyze data from discrimination experiments is to calculate a discrimination index. It is a simple ratio of the two response rates described above. Response rate in SΔ over response rate in SD is such an index (Rate in SΔ)/(Rate in SD) [ ＝ (Responses in SΔ/Time in SΔ)/ (Responses in SD/Time in SD)]. In most cases the discrimination index starts somewhere near a ratio of 1 (one) because the response rate is about the same in SD and SΔ the first few sessions. However, after several sessions the discrimination index will begin to decrease and ideally will reach a value very close to zero when the response rate in SΔ is very low. Notice that a discrimination index does not have a dimension such as seconds or resp/min (it is a ratio of response rates, so the dimensions divide out). Thus, when the rat does not show evidence of discrimination, the index is near a value of 1.0, and for perfect discrimination performance, the index is near or at zero. 5. Cumulative records should be used to display what happened in this experiment. In our case you may just want to show one from the first session and one from one of the last sessions where the discrimination was well established. You can combine two or three cumulative records on the same sheet of paper and refer to them as one Figure. However, you should label each cumulative record separately, such as Session 1 or Session 6, or use desсrіptive labels depending on what you want to show. Remember that there has to be a scale on the figure showing cumulative records. 6. You can plot all individual latencies in a session by simply importing the data to Excel and selecting the column with latencies and prepare a bar graph. This can be used to show within-session trends. If several sessions are compared, which is the ideal, then you should make sure that the y-axes match (have the same maximal value) 7. Frequency distributions of latencies are useful ways to display what happens in an experiment. Select important sessions and show the frequency distributions of latencies from those sessions. You should not show them for all sessions. Latency data are to be found in the raw data files and these files can be imported to Excel fairly easily. 8. For the test session for Conditioned Reinforcement (chaining), it would be appropriate to show the cumulative record. But it should be supplemented with a calculation in the text of response rate on both levers in SD and SΔ and a comparison should be made to an earlier session before the test to show that the rate of pressing the left lever increased in SΔ in the conditioned reinforcement test. Also, you may want to compare the rate on the left lever to the rate in previous sessions before the conditioned reinforcement test. The rate of pressing the left lever should be higher in the test than before. It is up to you what sort of figures you put in your paper. The analyses mentioned above are recommendations. Adequate coverage can be done with maybe 4 Figures but a better coverage will be 5-7 Figures. In any case, your paper should at least have one Figure with cumulative records (from discrimination training or testing for conditioned reinforcement). You see them all the time in the lab, and we talk about them all the time. For all figures, it is important to remember that you must describe in the Result section what the data in a given Figure show. What are the trends, what happens as a function of training, what happens when you change an independent variable, how quickly does the behavior recover after extinction, etc. It is very important for experimental reports in your writing to interact with the data in the Figures. Refer to the Figures frequently in your text, point to specific numbers, and point to trends in the data but without repeating all or nearly all the numbers on the Figures in the text. Guide the reader through the results; the main idea is to use the actual results to support what you write. See the additional document: Figures for the Discrimination Experiment for examples of what the figures can look like. It is not expected that you prepare all of these figures for your report. But the report has to have at least one figure with cumulative records of lever pressing for different phases of the experiment and at least one figurer with overall data such as response rates in SD and SΔ.
