The reliability of bilateral cerebral laterality for word generation: who is left in the middle?

Although word production is considered a strongly left hemispheric lateralized function, its cerebral asymmetry varies among individuals. The most popular way of determining hemisphere dominance is to calculate a laterality index (LI) by comparing brain activity between the two hemispheres. Large LIs can readily be classified as left or right dominant, but there is no consensus on how to treat bilateral LIs indicating (near) symmetrical activity. The problem with interpreting very small LIs is perpetuated by a lack of reliability, stemming from the challenge of systematically investigating these uncommon cases (usually ≤ 10% of a sample). To address this gap, we performed two studies that investigated the reliability and across-methods generalizability of bilateral LIs obtained from functional transcranial Doppler sonography (fTCDS), an ultrasound-based approach that measures cerebral blood flow velocity. In Study 1, we compared reproducibility of bilateral LIs (n=35) and clearly lateralized LIs (n=32) during a letter verbal fluency fTCDS task across two sessions. While left-lateralized classifications were highly replicable (97% reproduced), poorer reproducibility was observed for bilateral classifications (51% reproduced). In fact, dichotomous left-right categorization yielded more reliable outcomes than assigning participants with bilateral LIs to a separate category (80% reproduced). Study 2 assessed whether small fTCDS asymmetry extended to other instruments for determining cerebral laterality (visual half field method and fMRI). Participants consistently classified as bilateral by fTCDS (n=18) also exhibited reduced group-level asymmetry in these other methods. Based on these results, we suggest that LIs reflect a combination of idiosyncratic cerebral asymmetry, state-dependent fluctuations, and measurement noise. Our findings also indicate that a subset of the population has a neural system for word production that is inherently weakly lateralized, although true hemispheric equivalence is likely extremely rare. Finally, we offer recommendations for classifying asymmetry in clinical and research contexts.