James J. Gibson’s ecological approach to perception and action is founded on the principle of reciprocity (Lombardo, 1987; Heft, 2001; Adolph and Kretch, 2015). At the ecological scale of an individual animal, the animal’s behavior is tailored with respect to its surrounding environment, whereas the environment provides behavioral opportunities to the animal. Mutually complementing each other, animals and their surrounding environments are integrated as a reciprocally interactive ecosystem.

To account for the reciprocal interaction between the animal and the environment, Gibson (1977, 1986/1979) introduced what he termed “affordances,” environmental properties taken with reference to animals. In this view, an animal encountering the surrounding environment perceives a layout of surfaces scaled relative to the animal’s anatomy and its action capabilities, rather than objective qualities (e.g., shape, size, texture, color, composition, mass, and motion) indifferent to the animal’s scale and action capabilities. For example, for humans, a flat, rigid, extended, and knee-high surface affords sitting. However, a specific chair may be sit-on-able for an adult, but only climb-on-able for a toddler. An object can have many different affordances, depending on the observer’s action capabilities and/or behavioral goals, for affordances are related, not only to the environment, but also to the observer. For example, a specific chair may also afford being used as a step stool (climb-on-able) or to bar a door. As Gibson (1986/1979) put it, “The affordances of the environment are what it offers the animal, what it provides or furnishes, either for good or ill… It implies the complementarity of the animal and the environment” (Gibson, 1986/1979, p. 127, italics original).

To perceive an affordance, an animal must detect the information specifying it. This information is available in the structured energy pattern ambient at any given point of observation, and is uniquely determined by the environmental layout. Gibson argued that, to detect such information, the animal must seek it actively. In the words of Gibson (1986), “We must perceive in order to move, but we must also move in order to perceive” (p. 223). In the case of vision, this involves not only using the eyes, but “the eyes in the head on the shoulders of a body that gets about” (Gibson, 1986/1979, p. 222). Whereas the available affordance information must be actively detected by the animal, reciprocally, the animal’s action is guided by perceptual information. As the animal moves, its motion uniquely transforms the ambient energy distribution in accordance with the changes in the environmental layout and the displacements of the moving observation point. This transforming energy pattern at a moving point of observation, or optic flow, in turn, is specific to the environmental layout and the animal’s movements that engendered it (Gibson, 1966, 1986/1979; Warren, 1998, 2006). For example, forward movement structures the optic flow such that all the optical elements radiate from a single point, the focus of expansion, corresponding to the actor’s movement direction. By regulating the direction of movement coincident with the focus of expansion, the animal can reach its intended target.

As noted above, the ambient optic array at a given point of observation is structured exclusively by objects and the layout of the environment. Therefore, the structure of the optic array is specific to the facts of the animal’s surroundings. When an observer occupies a point of observation, the portion of the ambient array sampled by the observer is confined to the field of view registered by her eyes (see Figure 7.1 in Gibson, 1986/1979). This field of view, which can be portrayed as an oval window, contains various optical structures, some of which correspond to the observer’s body parts, such as the nose, lips, cheek, and limbs. As the observer moves, for example, rotating her head, those structures corresponding to her body parts transform. Because the optical transformation is produced by observer movement, the patterns of transformation are specific to the movements that produced them. For Gibson, perception of the environment and perception of the self were inseparable, always occurring together (Reed, 1996). As Gibson remarked, “One perceives the environment and coperceives oneself” (p. 126).

In the optic array, information specific to the environment is called extero-specific and information specific to the observer is termed proprio-specific. Since Sherrington (1906), it has commonly been thought that self-perception is conveyed by information from mechanoreceptors in the muscles, tendons, and joints. For Gibson, however, awareness of self can also be gained by visual information. Gibson referred to this visually based awareness of self as visual kinesthesis.

For Gibson, the concept of affordance binds the reciprocal pairs (duals) of animal and environment, perception and action, (subjective) self and (objective) world, into a dynamic whole.¹ As Gibson (1986/1979) put it,

An extensive body of research has demonstrated that human observers are capable of perceiving a variety of affordances (see Fajen et al., 2008, for review). Whether environmental layouts are perceived in body-scaled terms (Warren, 1984; Mark, 1987; Warren and Whang, 1987; Carello et al., 1989; Wraga, 1999) or action capabilities (Oudejans et al., 1996; Cesari et al., 2003), as Gibson contended, research findings to date confirm that affordances, rather than physical properties, are what animals perceive.

Because of schizophrenia’s complexity and still unidentified etiology and pathogenesis, researchers disagree as to how this disorder should be conceptualized. The present diagnostic system is grounded in the reductionist paradigm of neuroscience and biological psychiatry. This approach aims to identify the underlying cognitive and neurobiological processes for each symptom or symptom group comprising the diverse psychopathology of schizophrenia (Persons, 1986; Cahill and Frith, 1996). Proponents of phenomenological psychiatry and philosophy argue that the reductionist approach fails to recognize the illness’s first-person dimensions (Lysaker and Lysaker, 2010). Sass and Parnas (2003), in particular, contend that the disparate psychopathological symptoms of schizophrenia may be manifestations of a single phenomenological core, i.e., a disturbance of “ipseity” (ipse in Latin meaning “self” or “itself”), also referred to as minimal self, basic self, core self, or proto-self, which accounts for the pre-reflective (i.e., non-conceptual) and primitive level of subjective awareness, i.e., sense of ownership (i.e., an awareness of being the source of phenomenal experiences) and sense of agency (i.e., an awareness of being the agent executing one’s own actions) (Gallagher, 2000; Sass and Parnas, 2003; Zahavi, 2005; Stanghellini, 2009; Fuchs, 2010; Parnas and Sass, 2010; Nelson et al., 2014). The sources of core self distortion are thought to be two mutually interdependent processes—hyperreflexivity and diminished self-affection. Hyperreflexivity refers to a form of exaggerated self-consciousness in which tacit aspects of oneself become objectified as if they were external objects; Diminished self-affection refers to a weakening sense of being a subject of awareness (Sass and Parnas, 2003).

An individual’s awareness of his own thoughts, actions, perceptions, feelings, or pains operates at a pre-reflective (i.e., direct, immediate, implicit, non-conceptual, non-inferential, or non-reflective) level. This awareness of selfhood, as Sass and Parnas contend (Parnas, 2000, 2003, 2011, 2012; Sass, 2003a,b, 2014; Sass and Parnas, 2003; Cermolacce et al., 2007; Parnas and Sass, 2010; Raballo et al., 2011; Nelson et al., 2014), can be so profoundly disturbed in patients with schizophrenia that the distinction between self and others is altered or even vanishes. The patient’s body, normally taken for granted, begins to feel strange and unfamiliar, inviting explicit attention. As self-observation increases, aspects of the self begin to separate or detach such that “one’s arms or legs, one’s face, the feelings in the mouth or throat, the orbital housing of the eyes—even one’s speaking, thinking, or feeling” become “objectified, alien, and apart, perhaps even like the possessions of some foreign being” (Sass and Parnas, 2007, p. 432). The alienated body and aspects of one’s feeling increasingly command the individual’s introspection and intense reflection. Because the “detached” body no longer mediates between the self and the world, the patient’s normal sense of immersion in the world fades away and the world loses its practical meaning. Ultimately, the schizophrenia patient takes on the characteristics of a soulless body or a disembodied spirit³ (Stanghellini, 2009).

Twenty-six community-dwelling schizophrenia outpatients (19 males and 7 females) and 23 healthy control participants (18 males and 5 females) participated in the study. Schizophrenia participants had been diagnosed in accordance with the DSM-IV diagnostic criteria for schizophrenia (American Psychiatric Association, 1994) and were following individualized medication regimes. Schizophrenia participants were recruited from local (vocational) rehabilitation centers and mental health clinics. The control participants, who had no known family history of mental health disorder, were recruited from the university community.

For the schizophrenia group, the mean age was 41.0 ± 9.1 years, the mean years of education were 12.9 ± 2.25 years, the mean duration of schizophrenia was 18.2 ± 8.5 years, and the mean age of onset was 23.2 ± 4.4 years. For the control group, the mean age was 42.0 ± 7.45 years and the mean years of education were 13.6 ± 1.88 years. The two groups matched for age, t(47) = 0.42, p > 0.05, and education, t(47) = 1.24, p > 0.05.

One of the core features of schizophrenia is impaired cognition (Fioravanti et al., 2005, 2012; Goff et al., 2011). To exclude the possible effect of impaired cognition on task performance, we assessed each participant’s cognitive impairment using the Korean adaptation (Kwon and Park, 1989) of the Mini Mental State Examination (MMSE) (Folstein et al., 1975). For the schizophrenia group, the mean MMSE score was 28.1 ± 1.38, with five minimum scores of 26. For the control group, it was 28.5 ± 0.90, a negligible difference, t(47) = 1.315, p > 0.05.

All participants had normal or corrected-to-normal vision. Participants received a nominal ($5) fee for their participation in the experiment.

The study was approved by the Keimyung University’s Institutional Review Board. After providing a complete description of the study to the participants, written informed consent was obtained in accordance with the Declaration of Helsinki.

Color images of 27 household items served as the stimuli for the present experiment and are listed in Table 1. All images were sized to 600 × 800 pixels and presented on a 15-inch laptop with a pixel resolution of 1024 H × 760 V. Examples are shown in Figure 1. The presentation of stimuli was controlled by DirectRT (Jarvis, 2012), which also recorded responses and measured accuracy and reaction times of the responses. Participants viewed the display binocularly at a distance of approximately 50 cm.

Each artifact exhibited its designed affordance clearly. Three pairs of affordances were used: (a) scoop-with/pierce-with; (b) pour-in-able/stretchable; (c) cut-able-with/mop-up-with.⁵ As shown in Table 1, the nine objects selected to evaluate each affordance pair were divided into three different classes (cf. Ye et al., 2009): O_aff1 had only the first affordance (e.g., scoop-with) but not the second affordance (e.g., pierce-with); O_aff2 had only the second affordance but not the first; O_aff1,2 had both affordances (e.g., scoop-with and pierce with).

The experiment consisted of three randomized blocks of 108 trials with each block divided by paired affordances (see Table 1). Each block was further divided into two half-blocks with each half-block consisting of nine images of objects having either one or both of the paired affordances, that is, three images of objects with the first affordance (O_aff1), three with the second affordance (O_aff2), and three having both affordances (O_aff1,2) (see Figure 1 for the images of objects that made up the cut-able-with/mop-up-with affordance pair block). Each object appeared twice in each half-block for a total of 18 trials.

Trials were initiated when the participant pressed the space bar, triggering the appearance of a fixation point in the center of the display for 1000 ms, followed by the image of an object presented on a white background. Upon appearance of the object, participants were asked to press as fast and accurately as possible one of the two response keys (left/right shift keys) corresponding to the object’s affordance.

The two response keys were counter-balanced across participants. Thus, for each pair of affordances, an affordance was assigned to one response key, whereas the other affordance was assigned to the other response key, which was repeated once for a total of 18 trials. Then, the same 18 trials were repeated once again, this time by reversing the response keys.⁶ Thus, the 36 randomized trials comprising each block were produced by a 3 (Object Class: O_aff1, O_aff2, O_aff1,2) × 3 (Object) × 2 (Order of Presentation) × 2 (Repetition) scheme.

Participants were tested individually. They were told that household items can be used to perform various functions other than their typical (designed) function. As an example, the experimenter demonstrated that a paper cup designed to contain liquid could be also serve as a candle holder. Using this example, participants were told that their task was to identify one of two actions that could be performed using the displayed object by pressing the corresponding response key.

After 18 trials, participants repeated the same procedure with the response keys reversed for the paired affordances. Prior to the initiation of each half-block, instructions were displayed in text on the computer screen informing participants about the target functions of the displayed objects and the matching response keys.

The pair of affordances examined in each block was explained verbally using the basic descriptions of Ye et al. (2009), but with a slight modification to fit to the Korean culture (see their Appendix A, pp. 213–214).⁷

A 24-trial stimulus–response compatibility task preceded the experiment to allow participants to become familiar with the experimental setup. Displays were made of a red circle (8 cm in diameter) that appeared either to the right or left from the center. For the first 12 trials participants were told to press the spatially congruent response keys (i.e., the right shift key to the right dot and the left shift key to the left dot) as fast and accurately as possible, whereas for the last 12 trials they were told to press the spatially incongruent response keys to the red dot.

A practice session was built using representative objects for each affordance pair block: a gourd dipper for scoop-with-able vs. a nail for pierce-with-able; a mug for pour-in-able vs. rubber bands for stretchable; and a plastic knife for cut-with-able vs. a towel for mop-with-able. The two objects of each affordance pair appeared twice in a practice set. Each practice set was repeated until the participant demonstrated that she fully understood the procedure. A 4-trial practice session preceded each block of the experiment and was repeated in the middle of the block after reversing the response keys. The arrangement of the response keys for each practice set matched that for the 18 trials administered in the ensuing half-block experiment.

Feedback was not provided during the experiment. After two 18-trial sets with a pair of affordances were completed, a short break was given before proceeding to the next block.

Performance of the two groups was compared in terms of accuracy and reaction time. Note that the 36 trials comprising each block were produced by a 3 (Object Class: O_aff1, O_aff2, O_aff1,2) × 3 (Object) × 2 (Order of Presentation) × 2 (Repetition) scheme. For each object class, there were three instances of objects. However, there were no systematic relationships among the corresponding objects across the three object classes. The object classes simply were composed of representative objects whose properties were classified as a certain affordance type. In addition, for the class of objects with both affordances (i.e., O_aff1,2), whichever way they were identified, they were identified accurately. This object class was excluded from the analyses. Thus, for both accuracy and reaction time analyses, responses were collapsed across object and results entered into a 2 (group) × 2 (Object Class: O_aff1, O_aff2) × 2 (Order of Presentation) mixed design analysis of variance (ANOVA) with Group as between-subjects factor, and Object Class and Order of Presentation as within-subjects factors.⁸

Figure 2 compares group performance in terms of mean percent correct (left panel) and mean reaction time (right panel) as a function of the paired affordances [scoop-with/pierce-with affordance pair (top); pour-in-able/stretchable affordance pair (middle); cut-able-with/mop-up-with affordance pair (bottom)]. Participants with schizophrenia made more errors and took longer to respond than controls. The results of the accuracy analysis for each of the paired affordances are presented first, followed by the results of the reaction time analysis.

Twenty-six community-dwelling schizophrenia outpatients (16 males and 10 females) and 23 healthy control participants (14 males and 9 females) participated in the study. None had participated in Experiment 1. Schizophrenia participants had been diagnosed in accordance with the DSM-IV diagnostic criteria for schizophrenia (American Psychiatric Association, 1994) and were following individual medication regimes. Schizophrenia participants were recruited from local (vocational) rehabilitation centers and mental health clinics. The control participants, who had no known family history of mental health disorder, were recruited from the university community.

For the schizophrenia group, the mean age was 39.1 ± 7.32 years, the mean years of education were 13.5 ± 1.94 years, the mean duration of schizophrenia was 15.8 ± 6.1 years, the mean age of onset was 23.5 ± 3.8 years, and the mean MMSE score was 28.3 ± 1.09 (with 8 minimum scores of 27). For the control group, the mean age was 38.6 ± 6.95 years, the mean years of education were 13.6 ± 1.88 years, and the mean MMSE score was 28.5 ± 1.09. The two groups matched for age, t(47) = -0.23, p > 0.05, education, t(47) = 1.92, p > 0.05, and MMSE, t(47) = 0.55, p > 0.05.

All participants had normal or corrected-to-normal vision. Participants received a nominal ($5) fee for their participation in the experiment.

The same apparatus and the same color images of 27 household items used in Experiment 1 were adopted.

The same three affordance pairs from Experiment 1 were used again. This time, each affordance pair block was repeated once for each separate affordance for a total of six randomized blocks of 36 trials each for O_{aff 1} and O_aff2. Under this arrangement, the target affordance was the signal, and the non-target affordance was the distractor. As in Experiment 1, each block was further divided into two half-blocks by reversing the response keys after 18 trials.

The procedure used in Experiment 1 was replicated with one exception. This time, participants were asked to press one response key if the displayed object had a specific affordance but the other response key if it did not, doing so as quickly and accurately as possible.

The same 24-trial stimulus–response compatibility task used in Experiment 1 preceded the experiment to familiarize participants with the experimental setup. After 18 trials, the same procedure was repeated with the response keys reversed. Prior to the initiation of each half-block, instructions were displayed in text on the computer screen informing of the target function of the displayed objects and the response keys matching the signal and the distractor, respectively. The same 4-trial practice session used in Experiment 1 preceded each half-block of the experiment.

Feedback was not provided during the experiment. After each block, a short break was given before proceeding to the next block.

The 36 trials comprising each block were produced by a 3 (Object Class: O_aff1, O_aff2, O_aff1,2) × 3 (Object) × 2 (Order of Presentation) × 2 (Repetition) scheme. Although there were three classes of objects used in each block, one (i.e., O_aff1) was the target whereas the other (O_aff2) served as a distractor. Thus, a hit or a correct rejection was coded as a correct response and a miss or a false alarm was coded as an incorrect response. As in Experiment 1 and for a later comparison with the results of Experiments 1 and 3, responses for the class of objects with both affordances (i.e., O_aff1,2) were deleted from the analyses. Finally, because participants identified objects with a different affordance in each block, it is possible under the present design to assess performance across different affordances directly. Thus, for both accuracy and reaction time analyses, responses were pulled together from all six blocks after collapsing across object and object class and results entered into a 2 (Group) × 2 (Order of Presentation) × 6 (Affordance) mixed design ANOVA with Group as between-subjects factor, and Order of Presentation and Affordance as within-subjects factors.

Figure 3 compares performance of the two groups in terms of mean percent correct (top panel) and mean reaction time (bottom panel) as a function of affordance. A mixed design ANOVA on percent correct revealed main effects of Group, F(1,47) = 32.74, p < 0.001, ηp2 = 0.41, and Affordance, F(5,235) = 10.99, p < 0.001, ηp2 = 0.19, which was qualified by a significant Group × Affordance interaction, F(5,235) = 5.17, p < 0.001, ηp2 = 0.10. A simple effects analysis demonstrated a group difference for the pierce-with affordance, F(1,47) = 6.34, p < 0.05; the pour-in-able affordance, F(1,47) = 13.61, p < 0.01; the stretchable affordance, F(1,47) = 7.43, p < 0.01; the cut-able-with affordance, F(1,47) = 24.80, p < 0.001; and the mop-up-with affordance, F(1,47) = 6.81, p < 0.05, respectively. However, the effect of Affordance was statistically significant only for schizophrenia participants, F(5,43) = 14.48, p < 0.001. Schizophrenia participants misidentified the objects with the cut-able-with affordance with particularly high frequency (68%), replicating a pattern observed in Experiment 1. The ANOVA also confirmed a main effect of Presentation Order, F(1,47) = 7.87, p < 0.01, ηp2 = 0.14 (Figure 4). Unlike Experiment 1 in which both groups’ accuracy decreased slightly in the second half-block, the two groups of the present experiment performed slight better in the second half block.

The ANOVA on reaction time demonstrated main effects of Group, F(1,47) = 20.91, p < 0.001, ηp2 = 0.31, Affordance, F(5,235) = 5.64, p < 0.01, ηp2 = 0.11, and Presentation Order, F(1,47) = 77.16, p < 0.001, ηp2 = 0.62. The ANOVA also confirmed significant interactions between Group and Order, F(1,47) = 5.79, p < 0.05, ηp2 = 0.11 (Figure 4), and between Affordance and Order, F(5,235) = 3.54, p < 0.01, ηp2 = 0.07. With respect to the Group × Order interaction, reaction times were reduced in the second half-block for both controls (0.80 s vs. 0.62 s) and participants with schizophrenia (1.28 s vs. 0.98 s). However, the reduction for participants with schizophrenia appears even steeper. With respect to the Order × Affordance interaction, difference in reaction times across the six affordance objects in the first half-block for both controls and schizophrenia participants, F(5,43) = 12.28, p < 0.001, became insignificant in the second half-block.

With overall mean accuracy of 84% in Experiment 1 and 85% in Experiment 2, respectively, the effect of task on the performance of the two schizophrenia groups appears negligible. The two control groups also performed similarly with mean accuracy of 95% in Experiment 1 and 96% in Experiment 2, respectively. However, schizophrenia participants in Experiment 2 were more accurate than those in Experiment 1 in identifying the scoop-with affordance (87% in Experiment 1 vs. 96% in Experiment 2), reaching a level comparable to that of controls (98%).

By contrast, the impact of task on reaction time was quite noticeable, particularly for schizophrenia participants. Although the yes/no task shortened reaction time by 0.09 s (0.75 s in Experiment 1 vs. 0.71 s in Experiment 2) for controls, this difference did not reach statistical significance. For the schizophrenia group, by contrast, the yes/no task shortened reaction time by 0.39 s [1.52 s in Experiment 1 vs. 1.13 s in Experiment 2], and the difference was statistically significant at the 0.001 level according to a Tukey post hoc test. A one-way ANOVA comparing the four groups’ reaction time data [two control groups of Experiments 1 and 2 and two schizophrenia groups of Experiments 1 and 2] confirmed a main effect of Group, F(3,94) = 21.43, p < 0.001, ηp2 = 0.41.

The results of Experiment 1 demonstrated the effect of Presentation Order through degraded accuracy (possibly a fatigue effect), but faster reaction time (possibly a practice effect) in the second-half block, not only for schizophrenia participants but also for controls. The same Order effect was observed in the present experiment for both controls and schizophrenia participants. This time, however, both groups performed more accurately and faster in the second half-block (Figure 4), although the improved performance of schizophrenia participants did not achieve that of controls. When performance of the two groups during the second-half block was compared using a 2 (Group) × 6 (Affordance) mixed design ANOVAs, the ANOVA on percent correct confirmed main effects of Group, F(1,47) = 31.21, p < 0.0001, ηp2 = 0.40, and Affordance, F(5,235) = 10.46, p < 0.0001, ηp2 = 0.18, and a significant Group × Affordance interaction, F(5,235) = 4.08, p < 0.01, ηp2 = 0.08, the same effects observed when the data from both blocks were analyzed. Interestingly, however, in the combined analysis, group difference disappeared only for the scoop-with affordance. This time, however, performance difference of the two groups for the stretchable affordance also became indistinguishable, F(1,47) = 3.67, p = 0.06. The ANOVA on reaction time, however, confirmed only a main effect of Group, F(1,47) = 17.84, p < 0.0001, ηp2 = 0.28. Schizophrenia participants took longer to respond (0.98 s) than controls (0.62 s).

Overall, the impact of the simplified yes/no task on accuracy was limited, although it facilitated the identification of one particular affordance, i.e., the scoop-with affordance. The impact on reaction time became more apparent. Still schizophrenia participants took longer to respond than controls, even considering a likely practice effect that further decreased reaction time in the second-half block. It appears, therefore, that a yes/no task facilitated schizophrenia participants’ performance to some degree, but not enough to undermine our contention that perceptual capacity to identify affordances is likely impaired in patients with schizophrenia.

With respect to the main effect of Affordance in which certain affordances, e.g., the scoop-with affordance, were better identified than other affordances, it is possible that the degree to which the three objects chosen for each affordance represent the target affordance in the present study may have been different. That is, some objects may have been better exemplars of a particular affordance than others (see Ye et al., 2009, for a discussion on this issue). Of interest is the fact that the cut-able-with affordance, which caused more errors than other affordances for both controls and participants with schizophrenia in Experiment 1, became non-distinct for controls but remained problematic for participants with schizophrenia in Experiment 2. Another possibility is that the degree to which human observers are attuned to each affordance may not be the same. Thus, certain affordances are more salient than others. More research is needed to corroborate these possibilities.

Experiment 1’s participants also participated in Experiment 3.

The same apparatus and stimuli used in the previous experiments were used.

The experiment was conducted in six randomized blocks with each block comprising a pair of physical properties. The paired properties were: (a) pink/right angle, (b) fabric/circular, (c) green/wooden-material. As in Experiment 2, each of these three blocks was repeated once for a separate physical property with the objects from the target property as signals, and the objects from the non-target property as distractors. Each property was represented by three objects for a total of six objects comprising each block. The objects chosen for each physical property are listed in Table 2. The six objects were presented twice producing a total of 12 trials, which were repeated after reversing the response keys. Thus, the 24 randomized trials constituting each block were produced by a 2 (Object Class: O_phys1, O_phys2) × 3 (Object) × 2 (Order of Presentation) × 2 (Repetition) scheme.

The procedure used in Experiment 2 was replicated with one exception. Participants were asked to press one response key if the displayed object had a specific physical property and the other response key if it did not as fast and accurately as possible.

As in the previous experiments, a 2-trial practice session preceded each block of the experiment and occurred in the middle of the block after reversing the response keys. The arrangement of the response keys of each practice set matched that of the 12 trials administered in the ensuing half-block experiment. The representative images that were used to produce practice trials in each block were pink roses (pink), a window (right angle), cloth (fabric), a coin (circle), a green bell pepper (green), and a wooden toy dog (wood).

After 12 trials, participants repeated the same procedure with the response keys reversed. Prior to the initiation of each half-block, instructions were displayed in text on the computer screen informing participants about the target property in the displayed objects and the response keys matching the signal and the distractor, respectively.

Feedback was not provided during the experiment. After two 12-trial sets with a physical property were completed, a short break was given before proceeding to the next block.

The 24 trials comprising each block were produced by a 2 (Object Class: O_phys1, O_phys2) × 3 (Object) × 2 (Order of Presentation) × 2 (Repetition) scheme. As in Experiment 2, of the two classes of objects constituting each block, one was the target whereas the other served as a distractor. Thus, a hit or a correct rejection was coded as a correct response and a miss or a false alarm was coded as an incorrect response. Again as in Experiment 2, responses were combined from all six blocks after collapsing across object and object class and results entered into a 2 (Group) × 2 (Order of Presentation) × 6 (Physical Property) mixed design ANOVA for accuracy and reaction time analyses, separately.

Mean percent correct (top panel) and mean reaction time (bottom panel) of two groups are presented as a function of physical property in Figure 5. Unlike the previous experiments, schizophrenia participants were extremely accurate in identifying objects’ physical properties with an overall mean of 95 %, a level comparable to that of controls (97%). Indeed, the ANOVA confirmed that this difference was not significant, F(1,47) = 1.69, p = 0.20. The Group × Physical Property interaction, however, reached significance, F(5,235) = 3.09, p < 0.05, ηp2 = 0.06. A simple effects analysis revealed that schizophrenia participants made more errors in identifying the fabric objects (92%), causing this interaction, F(1,47) = 8.30, p < 0.01. Further inspection of data revealed that it was the kitchen sponge that caused most confusion for schizophrenia participants. The mean percent correct of the two groups for each object used in the fabric objects condition is shown in Table 3. The kitchen sponge image used in the experiment showed two layers, with one that was made of a synthetic fabric-like material (see the middle right panel in Figure 1). Note that today most kitchen sponges, including the image that was used in the present experiments, are made from foamed plastic polymers. In hindsight, we should have used a more intuitive fabric object. Interestingly, the control group identified the kitchen sponge as a fabric object as accurately (96%) as the other two fabric objects. It is not clear, therefore, why the kitchen sponge was particularly problematic for schizophrenia participants.

By contrast, schizophrenia participants took longer to respond (0.86 s) than controls (0.60 s), a pattern replicating that observed in the previous experiments. An ANOVA on reaction time confirmed main effects of Group, F(1,47) = 23.63, p < 0.001, ηp2 = 0.34, Physical Property, F(5,235) = 3.39, p < 0.01, ηp2 = 0.07, and Presentation Order, F(1,47) = 18.78, p < 0.001, ηp2 = 0.29. The interaction between Physical Property and Order also reached statistical significance, F(5,235) = 3.32, p < 0.01, ηp2 = 0.07. The effect of Order was significant for the right angled objects, F(1,47) = 7.64, p < 0.01; the fabric objects, F(1,47) = 7.69, p < 0.01; and the circular objects, F(1,47) = 10.03, p < 0.01. However, the difference in reaction times across the six physical propertied objects in the first half-block for both controls and schizophrenia participants, F(5,43) = 3.88, p < 0.01, disappeared in the second half-block, F < 1, ns.

The results of Experiment 3 were straightforward. Schizophrenia participants performed as accurately as controls in identifying physical properties contained in the objects, but showed slower reaction times. However, schizophrenia participants responded more quickly than in the previous experiments. Figure 6 shows performance of the two groups across the three experiments. Improved performance by the schizophrenia participants in Experiment 3 is quite evident. Two one-way ANOVAs were conducted to compare the performance of the four groups participated in Experiments 2 and 3, one on percent correct and the other on reaction time. Both ANOVAs confirmed main effects of Group [F(3,94) = 26.60, p < 0.001, ηp2 = 0.46, for accuracy; F(3,94) = 17.97, p < 0.001, ηp2 = 0.36, for reaction time]. A Tukey post hoc test on percent correct revealed that performance of the Experiment 2 schizophrenia group differed significantly from that of the other three groups, all at the.001 level. Importantly, the Experiment 3 schizophrenia group did not differ from the two control groups of Experiments 2 and 3. The same test on reaction time failed to find a significant difference between the two control groups, but found significant differences between the Experiment 2 schizophrenia group and the other three groups, all at the 0.001 level. Importantly, the same test failed to reveal a significant difference between the Experiment 3 schizophrenia group and the Experiment 2 control group.

Still, schizophrenia participants responded slower than controls. However, the level of performance demonstrated by schizophrenia participants (0.86 s) was comparable to that of the schizophrenia patients group (0.84 s) reported in Sevos et al. (2013). Interestingly, their control group performed comparably to that demonstrated by the controls of the present experiment (0.61 s vs. 0.60 s). Also notable in the Sevos et al. (2013) study was equally accurate level of performance by their two groups (91% controls and 90% patients), a similar pattern observed in the present experiment. Thus, the longer reaction times may have been the result of the impaired reaction time mechanism in schizophrenia patients, consistent with existing literature (Gale and Holzman, 2000, for review).

Experiment 3 was conducted to examine whether schizophrenia participants’ performance in the previous experiments was somehow affected by visual processing impairments of schizophrenia reported in the literature (Butler et al., 2008; Silverstein et al., 2015). For this purpose, we used the same color images of the objects used in the previous experiments. Accurate identification of the physical properties of the objects demonstrated by schizophrenia participants in the present experiment is assuring providing a strong basis to rule out possible influence of the reported visual deficits in schizophrenia on the task employed in the present study. Taken together, the degraded performance demonstrated by the schizophrenia participants in the two previous experiments is likely due to an impaired capacity to perceive affordances.