Calculate the suppression ratio from pre-CS baseline responses and responses during the CS to measure the strength of a conditioned emotional response.
Suppression Ratio Formula
- SR = suppression ratio, a number from 0 to 1
- CS = number of responses made during the conditioned stimulus
- PreCS = number of responses made during the baseline period immediately before the conditioned stimulus, matched in length to the CS
This is the measure introduced by Estes and Skinner and formalized by Annau and Kamin. It is often written as A / (A + B), where A is responding during the CS and B is responding before it. A value of 0 means the animal stopped responding entirely while the CS was on, which indicates strong conditioned fear. A value of 0.5 means responding was identical in both periods, which indicates no conditioning. Values above 0.5 mean responding sped up during the CS.
When the baseline period and the CS period are not the same length, the raw counts are not comparable and the ratio has to be rate corrected first:
- Dcs = duration of the CS period
- Dpre = duration of the pre-CS baseline period, in the same unit as Dcs
The calculator applies this correction when you open the durations panel and enter two different period lengths. If you leave the panel closed, it assumes the two periods were equal and uses the raw counts.
To find how many responses during the CS would produce a specific ratio, the formula is rearranged:
The multi-trial mode computes the ratio separately for each trial you enter, then reports the mean, the standard deviation, the standard error, and the change from the first trial to the last. That last figure is what tells you whether suppression is being acquired, which is a falling ratio, or extinguished, which is a rising ratio.
Interpreting Suppression Ratio Values
The suppression ratio is not linear with respect to how much responding was lost. Halving the ratio from 0.4 to 0.2 is a much larger behavioral change than moving from 0.5 to 0.45. The table below converts each ratio into the equivalent response rate and percent suppression so you can see what a given value actually represents.
| Suppression ratio | CS responding as % of baseline | Percent suppression | Reading |
|---|---|---|---|
| 0.00 | 0% | 100% | Complete suppression, asymptotic fear |
| 0.05 | 5.3% | 94.7% | Near complete suppression |
| 0.10 | 11.1% | 88.9% | Strong suppression |
| 0.20 | 25.0% | 75.0% | Moderate to strong suppression |
| 0.30 | 42.9% | 57.1% | Moderate suppression, below asymptote |
| 0.40 | 66.7% | 33.3% | Weak suppression, early acquisition |
| 0.50 | 100% | 0% | No conditioning shown |
| 0.60 | 150% | -50% | Facilitation, responding increased |
Reported group means in conditioned suppression studies usually sit near 0.5 before training and fall toward 0.0 to 0.1 after several CS and shock pairings, with stronger shock intensities reaching asymptote in fewer trials. Extinction runs the same numbers back toward 0.5.
Other Ways the Same Data Gets Reported
Papers do not all use the same index, and a value of 0.2 in one paper is not comparable to a value of 0.2 in another if the measures differ. Use this table to convert what you have or to check which index a paper used.
| Measure | Formula | No effect | Complete suppression |
|---|---|---|---|
| Suppression ratio | CS / (CS + PreCS) | 0.5 | 0 |
| Percent suppression | 100 * (PreCS – CS) / PreCS | 0% | 100% |
| Simple rate ratio | CS / PreCS | 1 | 0 |
| Difference score | PreCS – CS | 0 | equal to PreCS |
| Log ratio | log10((CS + 1) / (PreCS + 1)) | 0 | negative |
The suppression ratio is preferred over the difference score and the simple rate ratio because it is bounded between 0 and 1 and because it stays defined when CS responding falls to zero. The simple rate ratio and percent suppression both break down when a subject has a baseline of zero, since dividing by that baseline is undefined.
Example Problems
Example 1: equal periods. A rat presses a lever 40 times during the 60 second period before the tone and 8 times during the 60 second tone. The total across both periods is 8 + 40 = 48. The suppression ratio is 8 / 48 = 0.167. Responding during the CS was 20% of baseline, so percent suppression is 80%. That is moderate to strong suppression, consistent with a partly trained CS.
Example 2: unequal periods. The baseline was recorded over 4 minutes and produced 60 responses, while the CS lasted 1 minute and produced 9 responses. Using the raw counts would give 9 / 69 = 0.13 and would badly overstate the fear, because the baseline period was four times longer. Rescale the baseline to the CS length first: 60 * (1 / 4) = 15 responses. The corrected ratio is 9 / (9 + 15) = 0.375, and percent suppression is (15 – 9) / 15 = 40%. That is weak suppression, not strong suppression.
Frequently Asked Questions
What does a suppression ratio of 0 mean? It means the subject made no responses at all while the conditioned stimulus was present, even though it was responding normally in the period just before. The behavior that was maintained by food reinforcement was completely suppressed by the conditioned emotional response, so 0 represents the strongest measurable fear on this scale. A ratio of 0.5 is the opposite endpoint of interest, since it means the CS produced no change in the rate of responding and provides no evidence of conditioning.
Why would a suppression ratio come out above 0.5? Values above 0.5 mean the subject responded faster during the CS than during the baseline. That happens when the CS signals something appetitive, such as food, rather than shock, and it also happens by chance on individual trials when response rates are low and variable. A single trial at 0.55 is usually noise. A group mean that sits reliably above 0.5 across trials indicates genuine facilitation and means you are no longer measuring conditioned suppression.
Do the pre-CS and CS periods have to be the same length? The standard formula assumes they are, and most protocols use a baseline window matched exactly to the CS duration for that reason. If your periods differ, do not feed the raw counts into the formula. Convert both to rates, or scale the baseline count by the CS duration divided by the baseline duration, and then apply the formula. The calculator does this for you when you open the durations panel and enter both lengths. One related case to watch for is a subject that made no responses in either period. The ratio is then undefined rather than 0, since both the numerator and the denominator are zero, and that usually points to a satiated subject or a failed operant baseline rather than extreme fear. Trials like that are normally dropped, and the multi-trial mode excludes them from the mean and reports how many were removed.