The passive perception of irrelevant actions can facilitate or interfere with the execution of one’s own actions, known as ‘automatic imitation’ (AI). For example, when one is pressing down on a button, reaction times (RTs) are faster when observing a finger depression (compatible action) and slower whilst observing a finger lift (incompatible action). This phenomenon has been attributed to the mirror neuron system and is thought to represent a laboratory model of spontaneous motor mimicry which facilitates social interactions. AI is typically reduced or absent when the observed movement is produced by a non-human agent. However, previous findings suggest that the magnitude of this ‘human bias’ can be modulated by top-down factors, such as attention and prior instructions regarding whether the movement was produced by a human or non-human agent. This thesis aimed to further examine the role of attention and belief regarding stimulus agency in automatic imitation. Participants were required to perform a pre-specified key press or release response to a diffuse yellow flash go signal. This response was either compatible or incompatible with the finger or object movement, which was presented simultaneously. AI was measured by subtracting compatible from incompatible RTs to calculate the compatibility effect. Experiments 1a, 1b, 2 and 7 focused on exploring the role of attention in AI. Experiment 1a revealed that the human bias is dependent on when the go signal occurs. AI was greater for the finger stimulus relative to the object stimulus when the go signal occurred during the movement, but not after the movement. It is suggested that attention to the movement is reduced when the go signal occurs after the movement. This implies that the human bias in AI is dependent on attention being directed towards the movement. Experiments 1b and 2 indicated that AI was removed if a visual dual task was added, but that AI remained and was greater when an auditory dual task was added. This indicates that AI was removed when the visual dual task competed for cognitive resources with action observation. The facilitation of AI when an auditory dual task was added suggests that the additional cognitive load may have occupied cognitive resources required for the inhibition of imitation. These findings highlight that AI is susceptible to attentional load, implying that AI is not a strongly automatic process. Experiment 7 explored whether the spread of attention modulates the magnitude of AI by comparing a ‘diffuse’ go signal to a ‘focused’ go signal which directed attention to the stimulus movement. Significantly larger AI effects were produced for the group of participants who saw the focused flash first, indicating that focusing attention on the spatial location of the movement increased AI, and furthermore that initially observing the focused flash ‘trained’ participants to pay attention to the stimulus movement in the diffuse flash condition. Experiments 3 and 4 examined why AI effects for non-human stimuli are more likely to be significant when trials are presented in separate blocks (e.g. human vs. non-human stimuli) as opposed to randomly mixed trials. It was hypothesised that this pattern of previous results could be due to less attention being drawn to stimulus differences when stimuli are presented separately as opposed to mixed with a block of trials. However, in both experiments, AI effects were present for the object stimulus in the group of participants who observed the block of finger trials first. This suggests that the prior observation of the finger movement caused a carry-over of human agency to the object stimulus. Experiments 5, 6, 8 and 9 directly explored the role of belief regarding stimulus agency in AI by instructing participants that the object movement was generated by a human finger movement. Experiments 5, 6 and 8 provided preliminary evidence that AI is affected by belief instructions, but the effects were weak or confounded by spatial stimulus-response compatibility (SRC) effects (i.e. compatibility effects based on spatial correspondence of the stimulus and response location). Experiment 9 was designed to differentiate imitative compatibility from SRC effects, thus providing a pure measure of imitative compatibility. Imitative compatibility was present for the object stimulus after the belief manipulation. This demonstrates that a human belief regarding stimulus agency of the object modulated imitative compatibility effects due to the top-down knowledge that the movement was human generated, and not due to increased attention and SRC effects. The presented work has provided multiple lines of evidence which demonstrate that so-called ‘automatic’ imitation effects are strongly susceptible to top-down influences, including attention and belief regarding stimulus agency. The current work could be used to evaluate top-down modulation of imitation in autistic populations, as it has been proposed that top-down modulation of the automatic imitation pathway may be atypical in autism.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21909号 / 情博第692号 / 新制||情||119(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 神谷 之康, 教授 熊田 孝恒, 教授 西野 恒 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DGAM