Training groups	Training activities	Visuospatial perspective-taking	Verbal Interaction
(a) Socialised altercentric block building (Fig. 3)	Two children will be seated next to each other, and a toy bear will be seated at 180°, 90°, or 270° angular difference to the line of sight of the children. The experimenter will present a model figure built from blocks. Children will be asked first to reproduce the figure based on their own visual frame. Then, they will be asked to adopt the visual frame of the toy bear and reproduce the figure. Note that the children will be encouraged to discuss the perspective of the toy bear when building the block-figure. Therefore, both visuospatial perspective-taking and verbal interaction will be trained in this group	✓	✓
(b) Parallel altercentric block building (Fig. 3)	This training group will have the same procedures as that in the (a) socialised altercentric block building group, except children in this group are required to reproduce the block-figure without discussing the perspective of the figurine. Therefore, only visuospatial perspective-taking will be trained in this group	✓	✗
(c) Paired dialogic reading	The storytelling intervention by Ornaghi et al. [101] will be adapted in this training group. The experimenter will read the book series "The Adventures of Jack and Theo" [102] to the children. Children will listen to three stories in each session. After each story, children will be engaged in different language games and conversations. This training will minimize the use of the mental state lexicon to avoid perspective-taking and mentalising. No block-building activity will be involved in the training group. Therefore, only verbal interaction will be trained in this group	✗	✓