Sir, We have read the insightful letter by O’Callaghan and Hornberger (2016), wherein they propose a neurocomputational model for recent reports (Melloni et al. , 2016; O’Callaghan et al. , 2016) of social decision-making impairments in behavioural variant frontotemporal dementia (bvFTD). To follow up, here we outline a predictive, context-sensitive account of relevant mechanisms in bvFTD and other frontal disorders. First, we reanalyse our data (Melloni et al. , 2016) to characterize the neuroanatomical foundation of short-term strategies during social bargaining. Second, we propose a neurocognitive model addressing extant results and motivating predictions for further research on social negotiation. In our study (Melloni et al. , 2016), participants played as proposers in a validated repeated version of the ultimatum game, making offers on how to split a sum of money with another player. Such a role brings together self-centred and other-centred processes, as one’s own actions must be risk-assessed while predictions are made about the opponent’s upcoming decisions. Capturing these complexities, our model characterizes successful social negotiation in terms of three dynamic strategies. First, self-benefits must be maximized through an adaptation to self-perspective (ASP). Second, the opponent’s preferences and benefits must be acknowledged at each decisional step, via an adaptation to the other’s perspective (AOP). Third, and more crucially, proposers must integrate their own perspectives with those of others to successfully deploy a self-other integration strategy (SOIS). While ASP and AOP constitute short-term strategies (as they are driven by the immediately previous offer), SOIS is a long-term strategy that unfolds throughout successive instances of negotiation. As shown in Melloni et al., (2016), patients with bvFTD evinced normal short-term bargaining behaviour but they were impaired in long-term negotiation—which was also the case in patients with focal frontal damage. Upon reanalysing the data, we found that preserved ASP performance …