This paper compares an emergent interaction modality for the In-Vehicle Infotainment System (IVIS), i.e., gesturing on the steering wheel, with two more popular modalities in modern cars: touch in the central dashboard and speech. We conducted a between-subjects experiment with 20 participants for each modality to assess the interaction performance with the IVIS and the impact on the driving performance. Moreover, we compared the three modalities in terms of usability, subjective workload and emotional response. The results showed no statically significant differences between the three interaction modalities regarding the various indicators for the driving task performance, while significant differences were found in measures of IVIS interaction performance: users performed less interactions to complete the secondary tasks with the speech modality, while, in average, a lower task completion time was registered with the touch modality. The three interfaces were comparable in terms of perceived usability, mental workload and emotional response.