Heterogeneity of temporal trajectories of β-cell function parameters in the onset of dysglycemia is unknown. We aimed to characterize this heterogeneity over β-cell mass reduction (BCMR) by pancreatectomy. Non diabetic individuals (=70) underwent mixed-meal/oral glucose tolerance tests (MMTT/OGTT) and/or hyperglycemic clamps, before and after pancreatectomy (~50% tissue removal). We computed complementary functional parameters describing insulin secretion rate (ISR): glucose sensitivity (GS, the slope of the ISR vs glucose concentration relationship), rate sensitivity or first phase insulin secretion (RS), ISR at standardized fasting glucose (fISRstd, not from clamp), and potentiation (POT, the increased glucose-independent ISR along an oral test). Missing values were imputed via miss Forest algorithm. The reversed graphed embedding (RGE) framework was applied to changes in the parameters from OGTTs between pre- and post-surgery to identify groups of individuals with extreme combinations of the changes (archetypes). Differences between archetypes were assessed via Wilcoxon rank-sum test, (p threshold 0.01). We detected 4 archetypes, A1 to A4 (figure), with moderate stability (median adjusted Rand index=0.71). The archetypes were not different before surgery in terms of insulin sensitivity (IS, Matsuda index) and fasting and 2-hour glucose (fG and 2hG, respectively). In A1 β-cell functional parameters did not decrease. fG and 2hG did not increase. In A2 GS, fISRstd and RS, but not POT, strongly decreased from high values, accompanied by pre-surgery high ISR over the OGTT and average IS. Only fG increased (IS increased). In A3 only fISRstd decreased, while POT increased from low values. fG and 2hG increased. In A4 GS, RS and POT, but not fISRstd, decreased, with high pre-surgery POT. Pre-surgery mean glucose was high. fG and 2hG did not increase (due to increasing IS). Thus, we detected 4 archetypes for the trajectories of β-cell function parameters after BCMR. In A1 BCMR has no functional effect. In A2 the hypersecreting machinery is not preserved. In A3 BCMR decreases the fasting secretory tone, and after an oral stimulus increased potentiation tentatively compensates for this dysfunction. In A4 the β-cell response to an oral provocation is insufficient before BCMR (despite strong potentiation) and further reduces afterwards. This heterogeneity suggests different patterns of β-cell dysfunction induced by BCMR in the pathogenesis of diabetes.