Given the agreement of the results of experiments 2, 3, and 4 with physiological recordings of the primate visual system, one further experiment was conducted. Upside-down faces pose a rather serious challenge to recognition (Arnheim 1954; Kohler 1940; Scapinello and Yarmey 1970; Valentine and Bruce 1986; Yarmey 1971; Yin 1969; Young et al. 1987). These recognition decrements are generally accounted for by the inability of the visual system to recover configural information from these faces (Diamond and Carey 1986; Rhodes et al. 1993; Rock 1973, 1974; Tanaka and Farah 1993; Thompson 1980; Young et al. 1987). The single-cell recordings using upside-down faces give mixed results, depending upon the species involved. Perrett et al. (1982) found cells in the fundus of the superior temporal sulcus of rhesus monkeys that exhibit an equal magnitude of response to faces rotated 90° or to an inverted position. They argued that while inversion leads to impaired recognition in humans, the arboreal existence of monkeys might justify cells that code faces independent of rotation and inversion. The response latencies of these cells are, however, longer for inverted, compared with upright, faces (Perrett et al. 1988), a finding thought to reflect the underpinnings of longer response latencies of humans when processing inverted faces (Morton and Johnson 1989). Behaviorally, however, monkeys, like humans, have difficulty when dealing with upside-down faces (Perrett et al. 1988; Phelps and Roberts 1994).
It is important to note that these response measures are based on explicit processing of upside-down faces. Thus, there appears to be some discrepancy in the processing of upside-down faces between the properties of single cells, on the one hand, and behavioral responses, on the other hand. Recordings in sheep from an area homologous to the superior temporal sulcus in monkeys indicate that face-selective cells, which respond vigorously to upright sheep faces, are more or less quiescent in the presence of upside-down faces (Kendrick and Baldwin 1987).
Do the implicit representations set up by unattended upside-down distractors give rise to priming of upright attended targets in humans? In the light of our primate ancestry, it seems plausible that priming of upright human faces could be based on face-selective cells that are activated by a face regardless of its orientation. But our predominant experience with upright faces (Goldstein and Chance 1980) provides compelling grounds for humans to be like sheep, such that implicit and explicit measures of upside-down face processing are equivalent.
Was this article helpful?