This paper tries to introduce Freeman's theory on chaotic processes in
the brain as the background to knowledge construction. Experimental s
tudy in the field of olfactory perception shows that internally genera
ted neural processes prepare the organism to seek out inputs coming fr
om the external world. The system dynamics are nonlinear: organized st
ate appears when inhaling, and exhalation drives to chaotic states. Th
e world image is made by spatial patterns of neuronal network activity
which emerge from the chaos. Some epistemological consequences of tha
t model have special psychological significance: a. In an associative
memory system, each new memory alters the old ones when becoming tied
to them. This process takes place at the macroscopic level (neural net
works) and not at the microscopic one (isolated neurons). There is no
place into Freeman's model neither for ''grandmother neurons'', nor fo
r invariant memory stores. b. The lack of pattern invariance contrasts
with behavioral stimulus-response invariance. The same Freeman gives
an answer to this paradox in the interview included in this issue of P
sicothema. c. The sole ''information'' successfully transmitted from o
lfactory bulb to deeper areas of the brain is an active perceptual con
struct. d. If we can extend the former idea to other sensory systems i
n different species -as human species-, then we must conclude that the
re is no ''information processing mechanism'' in the brain (as suppose
d by cognitive functionalists) but just perceptual constructs based on
individual experiences. Information is constructed, not processed, th
rough deterministic chaotic dynamics. e. Both subject purposes and env
ironment impose the circumstances for neural activity. Therefore, inte
ntionality becomes the main key for knowledge construction.