Detection thresholds (DT) are used routinely to determine the odour-active compounds in foods. It is known that the composition of a food matrix, such as ethanol or sugar contents, has an impact on the release of flavour compounds, and thus on DTs. Hence, in the case of beers, DTs determined in alcoholic beers might not be appropriate for alcohol-free beers (AFB). Therefore, the aim of this study was to determine DTs for aroma compounds typically found in these products in an AFB-like matrix. The matrix was designed to match the sugar concentration and pH of an AFB brewed by a cold contact process. Six different concentrations of 24 volatile compounds were assessed using a three-alternative forced-choice procedure. This was carried out by ortho- and retronasal perception by either sniffing or tasting the samples, respectively. DTs were calculated using either a standard best estimated threshold (BET) method or by logistic regression, where the DT was estimated as the concentration at which 50% of the panellists could perceive the stimulus. Moreover, an algorithm for the removal of false positives was applied to adjust the panellists’ raw responses. Threshold values ranged from 0.011 μg/L for (E)-β-damascenone to 391 mg/L for acetic acid. Retronasal DTs were generally lower than orthonasal. Regarding the different calculation methods, DTs calculated by BET were higher than those from logistic regression. Removal of false positives also produced higher DTs than those from raw data. For example, orthonasal DTs for vanillin were 1,040 μg/L, 1,880 μg/L, 398 μg/L and 1,490 μg/L, for BET using raw and adjusted data, and logistic regression using raw and adjusted data, respectively. The results obtained in this study could be used to better understand the role of flavour compounds in AFB, as well as the effect of the calculation method chosen to prevent underestimated results.