You may want to truncate the normal distribution as shown here. They give an example with truncation between -2 and +2, which you can change to -1 and +1 in the obvious way.
Note, however, that the standard deviation of the truncated distribution will be much less than 1. If you want a standard deviation of 1 for the truncated normal distribution, then you will have to start with a different normal distribution (i.e., one that has a standard deviation >> 1 before truncation).
