Results
Initial results of the experiment are showed in Table 1.
| SCHEDULE OF REINFORCEMENT | FR-5 | VR-5 | FR-15 | VR-15 | FR-25 | VR-25 | FR-35 | VR-35 | FR-50 | VR-50 |
| Resistance to Extinction Over-train (Low) |
1.7 |
10.2 |
3.5 |
30.5 |
5.2 |
40.2 |
10.0 |
40.4 |
14.2 |
70.2 |
| Resistance to Extinction Over-train (Medium) |
4.5 |
10.0 |
15.2 |
25.0 |
5.2 |
20.5 |
5.4 |
50.2 |
10.3 |
65.0 |
| Resistance to extinction Over-train (High) |
3.4 |
10.8 |
5.5 |
20.2 |
5.8 |
30.5 |
5.2 |
40.0 |
5.7 |
80.0 |
| SPONTANEOUS RECOVERY |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
| Resistance to second extinction |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
Table 1. Experimental data
For these data, one-way ANOVA was used to compare means of resistance to extinction of the separate trials with regard to schedule type and time (FR-5 through VR-50 trials. The majority of schedule types showed significant difference as compared to other schedule types, and only several significance levels slightly exceeded confidence level of 0.05. These results show that the experiment was held in a correct manner and the rat responded differently to different reinforcement schedules.
The data were reformatted for multivariate SPSS analysis: schedule encoding was divided into two variables: ScheduleType (Fixed and Variable) and ScheduleRate (the number of repetitions ”“ 5, 15, 25, 35, 50). Fig. 1 and 2 show mean values of resistance to extinction and resistance to second extinction accordingly.
Figure 1. Mean values of resistance to extinction aggregated by schedule rate
Figure 2. Mean values of resistance to second extinction aggregated by schedule rate
In this analysis, there were several focus aspects: the interaction between the independent variables ”“ schedule and the level of over-training, and the impact of independent variables on the resistance to second extinction.
General results of univariate ANOVA can be written as follows. For resistance to extinction considered as dependent variable: ScheduleType ”“ F(1,4) = 10.149, p = .036<.05; ScheduleRate ”“ F(1,4) = 1.496, p = .358>.05, OverTrain ”“ F(1,4) = 1.143, p = .815 > .05, ScheduleType * ScheduleRate ”“ F(1,4) = 17.223, p = .001<.05, ScheduleType * OverTrain ”“ F(1,4) = .603, p = .570>.05, ScheduleRate * OverTrain ”“ F(1,4) = .625, p = .740 > .05.
The results of ANOVA show that the choice of the type of schedule (fixed versus variable) has a statistically significant effect on the resistance to extinction of the virtual rat. The schedule type variable which was measured initially as VRxx or FRxx was divided into two variables, schedule type (fixed versus variable) and schedule rate (the number of times the bar should be pressed for the trial). The results show that the interaction of these two variables is statistically significant with strong relationship as p=.001 (which is expectable), but there is no statistically significant interaction between OverTrain variable and either of the components of schedule (type and rate). The results for the resistance to second extinction showed that the virtual rat could not be extinguished twice, and, therefore, neither of the variables had a statistically significant impact on the resistance to second extinction.
Discussion
The analysis of between-subjects relationship of two independent variables, reinforcement schedule and overtraining level, showed that there was no statically significant interaction between these variables. The analysis of the experimental data disproved the hypothesis about the existence of second extinction as the resistance to second extinction was the same for all trials and was equal to zero. This basically means that the subject cannot be extinguished twice. However, the existence of resistance to second extinction has been considered in other research works (e.g. Moody, Sunsay & Bouton, 2006), and it was reasonable to analyze the presence of this phenomenon and variables affecting it in laboratory rats. Resistance to second extinction is also witnessed in human beings, and further consideration of this concept can provide important insight into the nature of learning and extinction. This experiment has one significant limitation: instead of a laboratory rat, the simulation was used. Although the simulation program is very effective for the purposes of exploring learning patterns, it still has own built-in algorithms which might not exactly reflect the behavior of real rats. Therefore, further insight into the relationships between reinforcement schedule and over-training time and their effects on resistance to second extinction can be gained after repeating this experiment on real rats.
Conclusion
The research of the mechanisms of extinction showed that the hypothesis about the existence of the secondary extinction of virtual rat was rejected, as there were no statistically significant differences between the values of resistance to secondary extinction for different schedule types, times and over-training levels. In addition to this, no statistically significant interaction between the type of reinforcement schedule and level of mouse over-training was identified. For all cases, the values of resistance to secondary extinction were zero. This result can be interpreted as the disproof of the existence of resistance to secondary extinction. However, the research was based on the results of laboratory simulation using virtual rat software, and therefore could be to a certain extent conditioned by the functionality and algorithms of the software. For precise conclusions on the nature of this phenomenon, it is recommended to repeat the experiment and measure the resistance to secondary extinction using magazine trained laboratory rats.