Probability is a number between 0 and 1 measuring how likely an event is. 0 = impossible, 1 = certain.

At NMA, probability appears everywhere: a neuron fires with probability p, a trial belongs to stimulus A with probability q, model parameters are drawn from a distribution.

Key rules:
• P(A or B) = P(A) + P(B) - P(A and B)
• P(A and B) = P(A) × P(B) if A and B are independent
• All probabilities of all outcomes sum to 1

import numpy as np

# Simulate 1000 coin flips
flips = np.random.choice(['H', 'T'], size=1000)
p_heads = np.mean(flips == 'H')
print(f'P(heads) ≈ {p_heads:.3f}')  # ≈ 0.500

🦌 Ilya says: NMA uses probability to describe UNCERTAINTY about neural responses. Neurons are noisy — stats handles that.