Temporospatial and semantic multisensory aspects contribute to bodily and spatial perception. An informative paradigm to study this is the Auditory Pinocchio Illusion, in which participants perceive an elongation of their finger upon vertically pulling their finger and hearing a concurrent upward pitch glissando. This arguably relies on anchoring (i.e., associating) the ecologically unrelated upward pitch glissando to the finger and allows to separately assess the role of semantic and multisensory contributions. However, what is needed for this anchoring to occur is unknown. In a first Experiment, we manipulated top-down attention to the finger upon which either an ascending or descending sound would be produced. In a second experiment, we compared how different bottom-up multisensory cues (arising from actions performed on the finger) concurrent to the ascending or descending pitch affected finger length perception. Participants either pulled, touched or stretched their finger. Through a perceptual judgment task of finger landmark localization and questionnaire ratings, we measured participants’ perceived finger length in both studies and separately assessed their sensory imagery skills. Our results show that attention alters finger length perception according to questionnaire ratings but not perceptual judgements, while concurrent multisensory signals similarly affect both measures. No relationship between these effects and participants’ sensory imagery was found. We suggest that while top-down associations between pitch and verticality are necessary and affect questionnaire ratings, they are not sufficient to affect perceptual judgements. Bottom-up somatosensory cues seem to be additionally needed to impact such judgements in this illusion.