First, the parameter "strain" can be measured via different techniques, for example tissue Doppler, speckle tracking, and even simple M-mode is able to measure a special kind of strain. Therefore the value of strain depends on how one measure the strain and what kind of strain is ment (natural strain, Lagrangschean strain, longitudinal, transversal, circumferencial strain etc.). The current standard is: measure strain with speckle tracking and measure the global longitudinal strain (GLPS). The cut off value for the mean GLPS of the complete LV is appr. -12 %. Decreased values (this means -11, -10 etc) are pathologic, increased values (this means -13, -14 etc.) is normal (beware of the algebraic minus sign). The strain only describes the intramyocardial movement of speckles, therefore no special strain values alone identify an ischemic heart. A pathologic strain may occur in ischemia, amyloidosis, myocarditis, HOCM etc.

The normal range of global longitudinal strain has been a fair topic of debate. It depends on what instrument  and software package you are using. -12% would actually be at least moderately decreased for us. Many consider -17 or -18 GLS (aka GLPS) to be the lower limit of normal for GE software, but that's not sensitive or specific for CAD. See the attached. In CAD vs. non-CAD populations there will be significantly different average GLS's, but that's about it. Following it serially for a decline would likely be better. As far as radial and circumferential strain, many don't use them because they aren't very reproducible. 

Article Myocardial Strain Analysis by 2-Dimensional Speckle Tracking...