The effect of housing rats in a stimulus rich versus stimulus poor environment on preference measured by sigmoid double demand curves