diff --git a/scipy/scipy.stats.fisher_exact/README.md b/scipy/scipy.stats.fisher_exact/README.md
index 9004790..adf5e5a 100644
--- a/scipy/scipy.stats.fisher_exact/README.md
+++ b/scipy/scipy.stats.fisher_exact/README.md
@@ -52,7 +52,7 @@ The input table is [[a, b], [c, d]].
 |a| b |
 |c|d|
 
-Where $N_A = a + c$ for the elements in group A and $N_B = b + d$ for the elements in group B.
+Where $N_A = a + c$ for the elements in group A (performance values of network A with $N_A$ as number of test pattern) and $N_B = b + d$ for the elements in group B (performance values of network B with $N_B$ as number of test pattern). 
 
 |||
 |---|---|
@@ -60,6 +60,14 @@ Where $N_A = a + c$ for the elements in group A and $N_B = b + d$ for the elemen
 |c|d|
 
 
+If network architectures are tested, typically, the same data set is used in both conditions and such $N = N_A = N_B$. 
+
+|||
+|---|---|
+|$N - c$| $N - d$ |
+|c|d|
+
+
 ## [Example](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.fisher_exact.html)
 
 ||Group A|Group B|