[1] 8w Activation and facilitation in the lexicon of schizophrenics

[1] 100w schizophrenicpatients of three different subtypes (chronic undifferentiated, schizoaffective and paranoid) and 15 normal control subjects completed a semantic priming version of the lexical decision task. Schizophrenic subjects demonstrated longer reaction times relative to normal controls; however, all of the subjects showed semantic facilitation effects. No significant differences in reaction times were found among the three schizophrenic subtype groups. These results suggest that initial sensory-perceptual processing is unimpaired in schizophrenia and thus provide evidence of intact automatic processing. The similarity in performance among the schizophrenic subtype groups refutes the theory of a paranoid/non-paranoid dimension at the initial level of automatic processing.

[1] 54w is generally considered to be a limited, non-specific cognitive resource which can be allocated to various processing tasks. The limitedcapacity theory of attention proposed by Posner and Snyder (1975) suggests that more demanding cognitive tasks required greater allocation of attentional resources while other tasks proceed automatically, with little cognitive effort and minimal focused attention.

[2] 94w The controlled process, also referred to as the limited-capacity attention mechanism, is characterized as slow-acting, operating only with conscious attention, and inhibiting the retrieval of unrelated information stored in the lexicon (the memory structure containing the basic units of meaning constituting the vocabulary). Conversely, the automatic spreading activation process is fast-acting, occurs without intention, and does not affect unrelated information in the lexicon (Schneider and Shiffrin, 1977;Shiffrin and Schneider, 1977). Building on Posner and Snyder's theory of controlled and automatic attention, Callaway and Naghdi (1982) proposed a two-process model of information processing in schizophrenia.

[3] 26w According to their model, the attentional deficit in schizophrenics is reflected in the controlled processes whereas the automatic processes are normal, or in some cases, supernormal.

[4] 70w They cite certain paradoxes in schizophrenic processing as support for their model. Although schizophrenics generally exhibit slower reaction times than normals on motor response tasks, they demonstrate faster alpha blocking and resistance to the self-stimulation effect in the auditory evoked potential. Callaway and Naghdi suggest that reaction timed tasks require conscious responses and are considered to be controlled tasks while involuntary responses, such as alpha blocking, are seen as automatic.

[5] 48w Cognitive theorists have addressed the issues of automatic and controlled attention in information processing in normal subjects. Posner and Snyder's (1975) attentional theories suggest that any input which shares a pathway in the memory system with a recently presented stimulus will be processed more quickly, resulting in facilitation.

[6] 83w A model of spreading activation developed by Collins and Loftus (1975) to explain how semantic memory is searched proposes that concepts and their related properties are represented as nodes in a semantic network. Activation of a concept node is thought to spread throughout the node matrix and to result in the partial activation of related property nodes. Because nodes in semantic memory are thought to be connected to their names in the lexicon, lexical activation is assumed to proceed in the same manner.

[7] 70w To clarify if access to word knowledge is the result of strategies and post-lexical processing or from the automatic activation of related nodes, priming paradigms have been utilized with normal populations (Donnerwerth-Nolan et al., 198 1;Meyer and Schvaneveldt, 1971). In the lexical decision task, two strings of letters are presented to the subject. These strings can be words or nonwords, related or unrelated, and can be presented sequentially or simultaneously.

[8] 67w If automatic activation plays a part in the retrieval of words from the lexicon, then words which are related will show a facilitation of recognition as indicated in reduced reaction times compared to control word conditions with unrelated words or non-words (Meyer and Schvaneveldt, 1971;Meyer et al., 1974;Brown et al., 1987). Semantic processing paradigms were also used by Kwapil et al. (1990) to investigate facilitation in schizophrenics.

[9] 39w Defining facilitation as the differences in reaction times for related and neutral word pairs, this study reported significantly greater facilitation for schizophrenics as compared with bipolar and normal controls. However, possible confounds to these results may have been introduced

[10] 17w because different word lists were used within groups so comparisons of individual differences could not be made.

[11] 215w Due to the proposed theoretical connection between attentional processes and association and the striking presence of thought disorder in schizophrenia, considerable attention has been given to the assessment of thought disorder. Manschreck et al. (1988) examined semantic priming in groups of thought disordered and non-thought disordered schizophrenics, unipolar affective patients and normal controls and found evidence suggesting greater facilitation in thought disordered schizophrenics compared to other subject groups. It was proposed that activation may be more pronounced in thought disordered schizophrenics and contributes to intrusions noted in schizophrenic speech. However, these findings should be viewed cautiously as they were based on a small number of subjects per group. Additionally, schizophrenic subjects were not delineated by diagnostic subtype. It appears that schizophrenics do not all manifest the same type of thought disorder symptoms (Kay, 1986) and that thought disorder can occur in a wide range of psychiatric disturbances (Marengo and Harrow, 1985;Carter, 1986). In his review of the information processing research, Magaro (198 1) concluded that, although thought disorders are present across the paranoid/non-paranoid dimension in schizophrenia, distinctive cognitive processing exists in both sub-types. Extending this idea, Chapin et al. (1988Chapin et al. ( , 1989) ) used the span of apprehension task and lexical decision task to evaluate automatic information processing in sub-types of schizophrenia.

[12] 20w Findings indicated that semantic facilitation was seen in all groups which suggests that automatic processes are not impaired in schizophrenia.

[13] 42w However, the results of both tasks indicated that the schizophrenic paranoid type did not score like the classic schizophrenic but more like the normal and affective control populations. These studies suggest heuristic value in further information processing research across differential diagnostic categories.

[14] 95w In the present study, a priming version of the lexical decision task was utilized to differentiate between the controlled and automatic attentional processes in three diagnostic categories (chronic undifferentiated schizophrenia, paranoid schizophrenia, and schizoaffective disorders) and one normal control group. It was hypothesized that both schizophrenics and control subjects would manifest automatic semantic facilitation and respond more quickly to semantically related word pairs. It was also hypothesized that a paranoid/ non-paranoid dimension would be evidenced such that paranoid schizophrenic subjects would demonstrate greater facilitation effects and faster overall reaction times as compared to non-paranoid schizophrenics.

[1] 79w A 4 x (2) repeated measures MANOVA was conducted on the response latencies and error rates for the lexical decision task. Diagnostic category was the between subject factor while the primetarget type was the within subject factor. To test for semantic facilitation, semantically related and unrelated prime-target pairs were utilized. As noted in Table 1, the main effect of prime type, semantically related versus unrelated, was significant for both error rate (F(1,56)= 13.53,p<O.O1) and reaction time (F(1,56)= 27.05, ~~0.01).

[2] 53w For all diagnostic groups, semantically related prime-target pairs led to fewer errors and faster responses when compared to the error rates and reaction times for unrelated prime-target pairs. Fig. 1 indicates that schizophrenics did not make significantly more errors than the normal control group. No interaction between diagnostic group and prime type emerged.

[3] 20w The overall mean reaction times were significantly different between groups (F(3,56) = 8.02, p < 0.01). Student Newman-Keuls analyses re-

[4] 116w TABLE I MAN0 VA results comparing semantically related and unrelated prime types on the lexical decision task Between subjects Diagnosis-error rate (ER) Diagnosis-reaction time (RT) Error-between (ER) Error-between (RT) Within subjects Type of prime (ER) Type of prime (RT) Diag x primetype (ER) Diag x primetype (RT) Error-within (ER) Error-within (RT) 3 1.16 ns 3 8.02 0.01 56 56 1 13.53 0.01 I 27.05 0.01 3 0.45 ns 3 0.54 ns 56 56 Type Of Prllnne vealed that the schizophrenic groups had significantly slower reaction times when compared with the normal controls (Fig. 2). However, they also showed the semantic facilitation effect in which related word pairs were responded to more quickly than unrelated word pairs.

[1] 33w The lexical decision task is particularly well-suited to investigate the automatic spreading activation of semantically related words in the lexicon because it avoids many of the potential confounds inherent in other cognitive paradigms.

[2] 43w It is a repeatedmeasures design which examines the relative difference between semantically related and unrelated prime-target word pairs. Since the critical comparison is within the individual, subjects serve as their own controls and the difficulties associated with appropriately matching control groups are eliminated.

[3] 24w Furthermore, in the lexical decision task, the mode of subject response remains the same and only the relationships between the primetarget pairs are manipulated.

[4] 41w Therefore, rather than relying on an overall deficit in performance, the significant effect is dependent on a specific, differential deficit for unrelated word pairs such that reaction times for associated word pairs are faster (Meyer and Schvaneveldt, 1971;Meyer et al., 1974).

[5] 15w The results of the present study suggest that initial sensory-perceptual processing is intact in schizophrenia.

[6] 92w Although schizophrenic subjects demonstrated longer overall reaction times, all of the schizophrenic groups, as well as the normal controls, manifested semantic priming. Response times to semantically related prime-target word pairs were faster than response times for unrelated word pairs. Although paranoid schizophrenics demonstrated faster reaction times than non-paranoid schizophrenics, they were not significantly different from other schizophrenic subtype groups. This is in contrast to the findings of previous studies (Margaro, 1981;Chapin et al., 1988;Chapin et al., 1989) and raises questions regarding the presence of a unique style of cognitive processing in paranoids.

[7] 21w Perhaps the hypervigilance typically seen in this subtype results in relatively shorter response latencies even while the underlying attentional deficit remains.

[8] 88w In summary, the presence of semantic facilitation in the lexical decision task supports the theory of Callaway and Naghdi (1982) that automatic attentional processing is intact in schizophrenics. These results not only contribute significantly to the understanding of the role of attention and semantic priming in information processing in schizophrenia but also suggest direction for further research. Future studies utilizing information processing paradigms to investigate the efficiency of controlled attentional processing in schizophrenia may provide evidence to clarify and define the nature of the attention deficit in schizophrenia.

[1] 85w The subects were 45 in-patients in a state regional psychiatric hospital and 15 normal controls. Sub-ject groups were as follows: 15 in-patients with schizophrenia, chronic undifferentiated type; 15 inpatients with schizophrenia, paranoid type; 15 in-patients with schizoaffective disorder; and 1.5 normal control non-patients. Psychiatric subjects were selected using the Research Diagnostic Criteria (RDC) classification checklist as well as DSM-III-R diagnostic criteria for determining inclusion in appropriate subtypes of schizophrenia. All subjects had at least four previous psychiatric hospitalizations and were receiving therapeutic doses of neuroleptics.

[2] 74w Controls were matched on the basis of age, level of current intellectual functioning as estimated by the Jastak Vocabulary Test (Jastak and Jastak, 1964), sex and race. All subjects were 45 years of age or younger, native English speakers, and had normal or corrected-to-normal vision. Subjects were excluded if they had a history of brain damage, neurological disorder, chronic medical illness, or a primary diagnosis of substance abuse as indicated in their medical records.

[3] 78w 4-field semi-automatic tachistoscope was utilized to control the timed exposure duration of stimuli. Two hand-held response buttons, attached to a quartz digital timer, were used to measure individual response latencies and indicate response type. The viewing distance from subjects' eyes to the target stimuli was 3 1 inches. Luminance intensity of the tachistoscope lamp was independently controlled from each target field and remained constant throughout the experiment at 70 for the fixation dot and 90 for the stimuli.

[4] 76w Stimuli consisted of 48 pairs of associated words, e.g., bread-butter and nurse-doctor, 48 pairs of unrelated words constructed from randomly matching 48 additional associated word pairs, e.g., rug-lock and carpet-key, 48 pairs of non-words, and 48 pairs of words and non-words (derived from a list sent by David Meyer; Meyer and Schvaneveldt, 1971). The stimuli letter strings were stencilled on 4 x 6 inch white, precisely calibrated stimuli cards manufactured for use with the Gerbrands tachistoscope.

[5] 20w Letter strings were handlettered on the cards using a lettering guide (Berol RapiDesign R-910) to create letters 5/16 inches high.

[6] 18w After signing the informed consent, subjects were administered the Jastak Vocabulary Test and their visual acuity was checked.

[7] 40w Subjects were then tested individually in a dimly lit room. Time was allowed for the subjects' eyes to adapt to the darkened room. The lexical decision task required subjects to determine if the target stimuli was a word or non-word.

[8] 100w Subjects were instructed to indicate 'word' responses by pressing the button in their preferred hand and signal 'non-word' responses by pressing the button in their nonpreferred hand. Accuracy and speed of response were emphasized in the instructions read to subjects. Following the presentation of a fixation dot at luminance 70 to 500 ms, two letter strings were presented simultaneously in the center of the tachistoscope screen at luminance 90 for a maximum of 3500 ms. Exposure was terminated upon response by the subjects. After completing ten practice trials, subjects completed 96 experimental trials. Response accuracy and reaction times were recorded.