Purpose: Persons with aphasia (PWAs) have been shown to have impaired attention skills that may interfere with their ability to successfully participate in speech and language therapy. Fluctuations in attention can be detected using physiological measures such as electroencephalography (EEG), but these measures can be impractical for clinical use. The primary purpose of this study was to investigate observable behavioral signs of attention as a means of measuring within-session fluctuations in attention by comparing behavioral ratings to physiological changes. Other aims were to understand the relationship between observable behaviors and task performance and to determine whether syntactic complexity influences behavioral attention. Method: Ten PWAs and 10 neurologically healthy adults underwent a sentence-reading task with 45 active and 45 passive sentences while video/audio and EEG data were recorded continuously. EEG data for each trial were classified into one of four levels of attention using a classification algorithm (Berka et al., 2004), and video/audio data were scored for accuracy and behavioral engagement by two trained speech-language pathologist students using a behavioral rating scale of inattention (Whyte et al., 1996). Results: Results showed that behavioral engagement was significantly correlated with task performance, with higher engagement scores associated with fewer errors. Behavioral engagement did not differ based on syntactic complexity for either group, but PWAs had significantly lower behavioral engagement scores when they were in lower/distracted states of physiologically measured vigilant attention. Conclusion: Behavioral observation may provide an alternative means of detecting clinically significant lapses in attention during aphasia therapy.