Using change scores in the independent variable in quadratic regression

I am running a quadratic regression in spss. The x variable has both positive and negative values (because it is a change score). When x is squared to create the quadratic term, all negative values become positive, obviously. But in this case (change scores) the negative and positive values on x actually means something and by squaring it, it changes its value. How do I manage this? For example: positive change participant gets a 2, Negative change participant gets a -2, when squared they both get a 4. I am losing valuable information here. Can anyone shed some light on this?