The alkaloid fraction (AF) from Psychotria nemorosa leaves was found able to inhibit both monoamine oxidase-A (MAO-A) and butyrylcholinesterase (BChE). However, the alkaloids responsible for the enzymatic modulation had not been identified. Our study aims indicating the multifunctional compounds using a metabolic profiling approach. As a first step, the alkaloids extraction was optimized. Leaves of P. nemorosa were extracted with methanol using an ultrasonic bath. Factors were screened by a Fractional Factorial Design. The Euclidean distances between the UPLC-DAD fingerprints and the blank injection were used as response, evidencing highly diverse extraction outcomes. Coupled to the plotting of effects per time point (effect fingerprints), thermolabile peaks could be indicated. Consequently, time and temperature were selected for further optimization, using a Central Composite Design. Finally, plant:solvent ratio was set at 1:50 (m/v), number of extractions at one, particle size at ≤ 180 μm, extraction time at 65 min and temperature at 45 °C, thus avoiding degradation. The fractionation step was performed using a solid-phase extraction method on silica cartridges, optimized by a Box-Behnken Design. Finally, sample concentration was set at 150 mg mL-1, 40% acetonitrile in dichloromethane was used as eluting solvent, and the eluting volume was 30 mL. Using the optimized method, 43 AFs were analyzed by UPLC-DAD and assayed for their BChE and MAO-A inhibitory potencies. To correlate chromatographic fingerprints and pharmacological activities, Orthogonal Projections to Latent Structure (O-PLS) modelling was employed and the regression coefficients of the model were analyzed and compared to the original fingerprints. Four peaks were indicated as multifunctional compounds. All compounds were purified and their structures were determined as azepino-indole derivatives, described for the first time in Psychotria species.