The temporal dominance of sensations (TDS) method has received particular attention in the food science industry due to its ability to capture the time–series evolution of multiple sensations during food tasting. Similarly, the temporal liking method is used to record changes in consumer preferences over time. The conjunctive use of these methods provides an effective framework for analyzing food taste and preference, making them valuable tools for product development, quality control, and consumer research. We employed the TDS and temporal liking data of strawberries that were recorded in our earlier study to estimate the temporal liking values from sensory changes. For this purpose, we used a reservoir network, a type of recurrent neural network suitable for time–series data. The trained models exhibited prediction accuracy of the determination coefficient as high as 0.676–0.993, with the median being 0.951. Further, we proposed two types of sensitivities of each sensory attribute toward the change in the temporal liking value. Elemental sensitivity indicates the degree that each sensory attribute influences the temporal liking. In the case of strawberries, the sweet attribute was the greatest contributor, followed by the attribute of fruity. The two least-contributing attributes were light and green. Interactive sensitivity indicates how each attribute affects the temporal liking in conjunction with other attributes. This sensitivity analysis revealed that the sweet attribute positively influenced the liking, whereas the green and light attributes impacted it negatively. The proposed methods offer a new approach to comprehensively analyze how the results of TDS are linked to those of the temporal liking method, serving as a step toward developing an alternative system to human panels.

Temporal dominance of sensations (TDS) methods are used to record temporally developing sensations while eating food samples. Results of TDS tasks are typically discussed using averages across multiple trials and panels, and few methods have been developed to analyze differences between individual trials. We defined a similarity index between two time-series responses of TDS tasks. This index adopts a dynamic level to determine the importance of the timing of attribute selection. With a small dynamic level, the index focuses on the duration for attributes to be selected rather than on the timing of the attribute selection. With a large dynamic level, the index focuses on the temporal similarity between two TDS tasks. We performed an outlier analysis based on the developed similarity index using the results of TDS tasks performed in an earlier study. Certain samples were categorized as outliers irrespective of the dynamic level, whereas the categorization of a few samples depended on the level. The similarity index developed in this study achieved individual analyses of TDS tasks, including outlier detection, and adds new analysis techniques to TDS methods.

The temporal dominance of sensations (TDS) method captures assessors’ real-time sensory experiences during food tasting, while the temporal liking (TL) method evaluates dynamic changes in food preferences or perceived deliciousness. These sensory evaluation tools are essential for understanding consumer preferences but are also resource-intensive processes in the food development cycle. In this study, we used reservoir computing, a machine learning technique well-suited for time-series data, to predict temporal changes in liking based on the temporal evolution of dominant sensations. While previous studies developed reservoir models for specific food brands, achieving cross-brand prediction—predicting the temporal liking of one brand using a model trained on other brands—is a critical step toward replacing human assessors. We applied this approach to coffee products, predicting temporal liking for a given brand from its TDS data using a model trained on three other brands. The average prediction error across all brands was approximately 10% of the maximum instantaneous liking scores, and the mean correlation coefficients between the observed and predicted temporal scores ranged from 0.79 to 0.85 across the four brands, demonstrating the model’s potential for cross-brand prediction. This approach offers a promising technique for reducing the costs of sensory evaluation and enhancing product development in the food industry.

The Temporal Dominance of Sensations (TDS) method allows for the real-time tracking of changes in multiple sensory attributes, such as taste, aroma, and texture, during food tasting. Over the past decade, it has become an essential tool in sensory evaluation, offering novel insights into temporal sensory perception. When combined with the Temporal Liking (TL) method, TDS enables the investigation of how sensory changes influence instantaneous liking. Existing methods in time-series sensory evaluation have not simultaneously achieved the following two key objectives: (1) predicting TL curves from TDS curves and (2) identifying shared sensory–liking synergies across samples. In this study, we address this gap by applying supervised non-negative matrix factorization, which enables both precise prediction and interpretable synergy extraction. This novel approach has the potential to extend the applicability of TDS analysis to broader sensory evaluation contexts. We validated the method using the data for strawberries recorded in an earlier study. Our model, utilizing three latent synergy components accounting for 94% of the data variation, accurately predicted the TL curves from TDS curves with a median RMSE of 0.36 in cross-validation, approximately 1/16 of the maximum TL score. Moreover, these synergy components were highly interpretable, suggesting some key factors that explain individual variations in sensory perception. These findings highlight the effectiveness of synergy analysis in time-series sensory evaluation, leading to deeper understanding of the connections between temporal sensory and liking responses.

The temporal dominance of sensations (TDS) and temporal liking (TL) methods offer complementary insights into the evolution of sensory and hedonic responses during food consumption. This study investigates the feasibility of predicting TL curves for food pairings from their TDS profiles using reservoir computing, a type of recurrent neural network. Participants evaluated eight samples—two crackers (plain, sesame), two spreads (peanut butter, strawberry jam), and their four binary combinations—performing both TDS and TL evaluations. This process yielded paired time-series data of TDS and TL curves. We trained various reservoir models under different conditions, including varying reservoir sizes (64, 128, 192, or 256 neurons) and the inclusion of auxiliary input dimensions, such as flags indicating the types of foods tasted. Our results show that models with minimal auxiliary inputs achieved the lowest root mean squared errors (RMSEs), with the best performance being an RMSE of 0.44 points on a 9-point liking scale between the observed and predicted TL curves. The ability to predict TL curves for food pairings holds some promise for reducing the need for extensive sensory evaluation, especially when a large number of food combinations are targeted.