Neutral theory: 

The vast majority of observed sequence differences between members of a population are neutral (or close to neutral). These differences can be fixed in the population through random genetic drift. Some mutations are strongly counter selected (this is why there are patterns of conserved residues). Only very seldom is a mutation under positive selection. 

The neutral theory does not say that all evolution is neutral and everything is only due to to genetic drift.

(Nearly neutral theory:  Even synonymous mutations do not lead to random composition but to codon bias.  Small negative selection might be sufficient to produce this bias. )

Note: the larger the population the better selection works, and the closer to neutral a mutation needs to be in order to be fixed by genetic drift. (If N*s<<1 the mutation behaves as neutral, and the fixation probability is 1/N; if N*s~1 then fixation probability is only about 2s, which is small, but seems to work.)

Is Evolution in humans only neutral? Does selection still play a role? E.g., here , the distribution of alleles that encode a protein presumably involved in brain development (here for the article, in case you are interested to read more, a similar case reported here), here for a comment that argues the haplotype frequencies might be due to drift and small founder populations and not reflect selection.