RPE may be useful for exercise testing and prescription in individuals with spinal cord injury (SCI), although the roles of differentiated central and peripheral fatigue during exercise are not clear. We aimed to model differentiated RPE responses during graded arm cycling in individuals with SCI and to describe their relationship to cardiorespiratory outcomes. Thirty-six individuals with SCI (13 paraplegia and 23 tetraplegia) completed a maximal graded arm cycling exercise test to volitional exhaustion (5 W·min paraplegia; 10 W·min tetraplegia). Participants were asked to report central RPE (CRPE) and peripheral RPE (PRPE) every minute using the Borg category ratio (CR10) scale until termination of exercise. Heart rate and breath-by-breath respiratory outcomes were collected throughout the exercise test. Ventilatory threshold (VT) was assessed using the ventilatory equivalents method. Cardiorespiratory indices increased linearly during graded arm exercise. By contrast, both CRPE and PRPE responses were best fit to a quadratic model with positively accelerating growth in individuals with paraplegia (P < 0.01) and tetraplegia (P < 0.05). PRPE developed faster than CRPE in individuals with tetraplegia (P < 0.01). Individuals with paraplegia had accelerated CRPE (P < 0.05) and PRPE (P < 0.05) responses compared with tetraplegia, but not when considering only individuals who reached VT. PRPE was higher than CPRE only in the late stages (80%-100% test duration; P < 0.05) in both groups when only considering individuals who reached VT. PRPE develops faster than CRPE in individuals with tetraplegia in a nonlinear fashion, despite linear increases in cardiorespiratory responses during graded arm cycling. Although there is promise to use differentiated RPE for exercise testing and prescription within the SCI population, our results indicate that there are differences in how individuals with tetraplegia perceive peripheral versus central exertion.