Participants in the SE group will be also receiving two types of SE training tasks in each session, to match with the ones in the WM training. There are two mini-games for training.

To match with the spatial-visual WM training, participants in the SE group will be asked to first look at a picture that depicts a daily social scenario (i.e., a child who fell down and hurt himself), and then to choose an emotion that the child in the picture may feel from four options of facial emotions displayed on the screen. The other mini-game, with the aim to match with the auditory WM training, will first tell a short story of a social scenario (i.e. "David accidentally fell down and got hurt), and then will ask the child to choose the emotion that the character in the story may feel from the four options displayed on the screen. Each mini game lasts for about 7 minutes in each training session.

A computerized task that examined participants' abilities in discriminating discrepancy between different durations of time (Gooch et al., 2011). A threealternative forced-choice (oddball) paradigm was used. On each trial, the child heard three 1000Hz tones, two of which were 1200ms long and a roving target which was at a different length (400ms, 700ms, and 1100ms - each repeated nine times). The child was required to choose which interval was "different" from the other two. Six easy trials of 100ms comparing with 1200ms were incorporated randomly during the experiment to evaluate the attention level of the participants.

A computerized task that examined participants' abilities in discriminating discrepancy between different durations of time (Gooch et al., 2011). A threealternative forced-choice (oddball) paradigm was used. On each trial, the child heard three 1000Hz tones, two of which were 1200ms long and a roving target which was at a different length (400ms, 700ms, and 1100ms - each repeated nine times). The child was required to choose which interval was "different" from the other two. Six easy trials of 100ms comparing with 1200ms were incorporated randomly during the experiment to evaluate the attention level of the participants.

A computerized lightbulb game was used to measure participants' abilities in recognizing and producing a designated period. The task was adapted from a similar task developed by Gooch and colleagues (2011). It was presented in a computer-game-like format where the child was asked to look at the red lightbulb that would appear on the screen and remember how long it had been left on (3s, 5s, 6s, 9s, 12s, and 15s). After, the child was asked to reproduce for the same length of duration. Immediately following this presentation the screen went blank and the participant was asked to "turn the yellow lightbulb on and leave it on" (use "ENTER" to turn on) for the same amount of time as it had been for the red lightbulb, and then use "SPACE" to turn off the light. Each target duration is presented twice in a randomized order for a total of 12 trials.

A computerized lightbulb game was used to measure participants' abilities in recognizing and producing a designated period. The task was adapted from a similar task developed by Gooch and colleagues (2011). It was presented in a computer-game-like format where the child was asked to look at the red lightbulb that would appear on the screen and remember how long it had been left on (3s, 5s, 6s, 9s, 12s, and 15s). After, the child was asked to reproduce for the same length of duration. Immediately following this presentation the screen went blank and the participant was asked to "turn the yellow lightbulb on and leave it on" (use "ENTER" to turn on) for the same amount of time as it had been for the red lightbulb, and then use "SPACE" to turn off the light. Each target duration is presented twice in a randomized order for a total of 12 trials.

Subjects were required to indicate the end of 3, 5, 12, and 20-s intervals (each produced twice) by saying "STOP" to indicate the end of the temporal duration.

Subjects were required to indicate the end of 3, 5, 12, and 20-s intervals (each produced twice) by saying "STOP" to indicate the end of the temporal duration.

This task was developed by Garner et al. (1994) to measure children's ability to infer emotions from situational cues, and similar tasks have been used among preschoolers (Beaudoin et al. 2020; Gallant et al. 2020). Social situations commonly encountered by children were presented in pictures, with the facial expressions of the target characters left out (i.e., blank faces were shown). The children were asked to identify the feeling of the target character by either naming the emotion or pointing to the corresponding facial expression out of four options (happy, sad, angry, and afraid) given on the stimulus sheet. There were two sample items followed by 13 test items.

This task was developed by Garner et al. (1994) to measure children's ability to infer emotions from situational cues, and similar tasks have been used among preschoolers (Beaudoin et al. 2020; Gallant et al. 2020). Social situations commonly encountered by children were presented in pictures, with the facial expressions of the target characters left out (i.e., blank faces were shown). The children were asked to identify the feeling of the target character by either naming the emotion or pointing to the corresponding facial expression out of four options (happy, sad, angry, and afraid) given on the stimulus sheet. There were two sample items followed by 13 test items.

Children's ToM will be measured by the ToM scale developed by Wellman and Liu (2004). The seven-item scale was designed for the experimental to adminster in order according to the difficulty of the task, with the Diverse Desire task as the easiest, and the Real-Apparent Emotion task as the most difficult. Participants will receive one point for each task passed successfully(range: 0-7). A higher score indicates a better ability of ToM.

Children's ToM will be measured by the ToM scale developed by Wellman and Liu (2004). The seven-item scale was designed for the experimental to adminster in order according to the difficulty of the task, with the Diverse Desire task as the easiest, and the Real-Apparent Emotion task as the most difficult. Participants will receive one point for each task passed successfully (range: 0-7).A higher score indicates a better ability of ToM.

The Working Memory Subscale from the BRIEF-P is used to measure participants' daily behavioural manifestation of working memory reported by parents. The minimum raw score is 0 and maximum score is 17. A higher score indicates more severe deficits in working memory.

The Working Memory Subscale from the BRIEF-P is used to measure participants' daily behavioural manifestation of working memory reported by parents. The minimum raw score is 0 and maximum score is 17. A higher score indicates more severe deficits in working memory.

The Hong Kong Reading Ability Screening Test for Preschool Children (RAST-K) will be used to measure participants' Chinese word reading ability. The maximum score is 55 and the minimum is 0. The higher score indicates a better Chinese word reading ability.

The Hong Kong Reading Ability Screening Test for Preschool Children (RAST-K) will be used to measure participants' Chinese word reading ability. The maximum score is 55 and the minimum is 0. The higher score indicates a better Chinese word reading ability.

An adapted English word reading assessment designed by the experimenters will be implemented to measure participants' English word reading ability. The maximum score is 55, and the minimum score is 0. A higher score indicates a higher ability of English word reading.

An adapted English word reading assessment designed by the experimenters will be implemented to measure participants' English word reading ability. The maximum score is 55, and the minimum score is 0. A higher score indicates a higher ability of English word reading.

The numeration subscale from Keymath-3 Diagnostic Assessment (Connolly, 2014) will be used to measure their number skills. The maximum score is 49 and the minimum score is 0. A higher score indicates a higher ability in numeration.

The numeration subscale from Keymath-3 Diagnostic Assessment (Connolly, 2014) will be used to measure their number skills. The maximum score is 49 and the minimum score is 0. A higher score indicates a higher ability in numeration.