Effective athlete preparation requires an understanding of the dose-response relationship between training load (TL) and fatigue. Acute overreaching (OR) related to fatigue may present as muscle soreness and changes to mood(1). Non-functional overreaching (NFOR) may present as disturbed sleep(2) and inhibited psychomotor speed(3). Understanding the link between TL and fatigue can allow NFOR to be avoided(1). Inherent difficulties in measuring the latent physiological responses to exercise means multiple measures of fatigue may be required to understand the effects of a given TL. Exploratory factor analysis (EFA) may provide insight to which measured variables best reflect these latent responses. Mixed martial arts (MMA) is a cross-discipline combat sport requiring multiple forms of training. Previous research has demonstrated that MMA training does not feature weekly changes in TL or fatigue, making identification of a dose-repose relationship difficult(4). Therefore the aim of this study was to determine the relationship between the latent variables related to TL and the latent variables related to OR and NFOR in MMA training using EFA(5). Following institutional ethical approval n = 14 human MMA participants (age = 22 ± 4.4 years; habitual mass = 71.3 ± 7.7 kg; stature = 171 ± 9.9 cm) took part in this study for 8 consecutive weeks. Participant’s TL (sessional rating of perceived exertion, strain and monotony) was recorded after every training session(6). At the end of each day participants recorded soreness using a CR10 scale for the following body regions: head and neck; shoulders and arms; upper torso; lower torso; legs. Fatigue score was measured via short questionnaire of fatigue at the end of each day(1). The following variables were measured as central fatigue proxies at the start of each week: overall sleep quality and sleep efficiency via Pittsburgh Sleep Quality Index (PSQI)(2); choice reaction time (CRT) and CRT correct count via Deary-Liewald CRT test(3). EFA was completed in JASP 0.14.1.0 following established protocols(5). Relationships between factors were assessed via Pearson’s correlation coefficient. 112 observations of 15 variables were included in the model which was found to have acceptable RMSEA = .069 and good TLI = .969 whilst explaining 64% of total variance in the data. Three factors (Figure 1) were identified via scree plot inflection point and an acceptable number (max=5) of factor loadings ≥±.30. Factor 1 may be labelled ‘Load’ and contains relationships between each TL variable apart from strain. Factor 2 may be labelled ‘Fatigue and Soreness’, containing relationships between body region soreness (apart from legs) and fatigue score. Factor 3 may be labelled ‘Central Effects’, containing relationships between sleep quality, sleep efficiency and CRT correct count. Factor 1 was found to have a moderate, negative relationship with Factor 3 (r= -.431), and a trivial positive relationship with Factor 2 (r=.095). Factor 2 also had a trivial, negative relationship with Factor 3 (r= -.021). The three factor model identified suggests static MMA TL causes central fatigue effects that may be indicative of the onset of NFOR in this population, despite an absence of acute changes in OR measures(4).