Application of the Self-Generation Effect 1 The Application of the Self-Generation Effect to the Learning of Blissymbols by persons with a Severe Aphasia

This study investigated the application of the self-generation effect to enhance the recognition and retention of Blissymbols in persons with severe aphasia. A 2X2X3 factorial design of two treatment types (self-generation and non-generation) was used to teach two sets of Blissymbols. These were administered during three training days, between which were withdrawal periods of one day and seven days. Recognition and retention probes were administered at intervals during the training. ANOVA analysis showed that the self-generation treatment produced no immediate recognition advantage; however, better retention of symbol recognition may have occurred over time. Hence, the potential application of the self-generation effect in enhancing the retention of Blissymbols in persons with severe aphasia may warrant further investigation. Application of the Self-Generation Effect 2 The self-generation effect (SGE) refers to the finding of superior retention and recall for stimuli constructed or generated by a normal, healthy adult. Memory for stimuli such as words, numbers and pictures were found to be enhanced by the extent to which the individual was involved in its construction (Ghatala, 1981; Jacoby, 1978; McFarland, Frey & Rhodes, 1980; Peynircioglu, 1989; Slamecka & Graf, 1978; Tyler, Hertel, McCallum & Ellis, 1979). Although several studies have shown that individuals with severe aphasia can acquire graphic symbol systems like Blissymbols, further investigation is required to identify the methods that best improve the acquisition, retention and communicative use of such symbol systems (Beck & Fritz, 1998; Bertoni, Stoffel & Weniger, 1991; Koul, Corwin & Hayes, 2005; Koul & Harding, 1998; Koul & Lloyd, 1998; McCall, Shelton, Weinrich & Cox, 2000; Weinrich, Shelton, McCall & Cox, 1997). This study proposes that the SGE may provide one such option for improving the recognition and retention of Blissymbols. Research into the SGE has found it to be a robust phenomenon of memory that emerges during free recall and recognition testing (Slamecka & Graf, 1978). Different types of construction or generation methods have also successfully elicited the SGE through deriving opposites, synonyms, and rhyming words; solving mathematical problems; and completing sentences. The SGE does not appear to be influenced by the type of generation method applied; rather, it emerges during recall or recognition testing when there has been effortful processing and encoding of the to-be-remembered stimuli’s attributes at the time of generation (Jacoby, 1978; Ghatala, 1981; Tyler et al., 1979). Application of the Self-Generation Effect 3 Jacoby (1978) tested the benefits of SGE for word stimuli generated via problem solving. In his first experiment, 18 participants were required to either read or construct member pairs of related words. The subjects were required to generate a word related to a cue word using the letters provided as clues (e.g., dog b_ n_ bone) or to read out the word pairs provided (e.g., sun – shine). Jacoby (1978) reported that subjects recalled more generated words than read words during testing. He attributed this superior recall for the generated word stimuli to SGE. Using healthy college students, Tyler et al. (1979) compared recall for word stimuli derived from a simple sentence completion task to a more challenging anagram task. In the sentence completion task, the words to be recalled were selected from two words provided to fill in a blank in a sentence (e.g., The girl was awakened by a frightening...... dream or apple). In the anagram task, the participants unscrambled a series of letters to identify a word (e.g., dortoc unscrambled to doctor). The data of Tyler et al. (1979) showed superior recall for word stimuli generated by the more challenging anagram task than by the sentence completion task. Similarly, McFarland, Frey & Rhodes (1980) also showed that the self-generation factor appeared to increase recall for generated word stimuli. The subjects normal, healthy college students either generated a word that rhymed with a given word or determined whether a pair of given words rhymed. Again, a stronger SGE was elicited for the generated items compared with the nongenerated items. Similarly, other types of stimuli, such as pictures, have also successfully elicited the SGE (Kinjo & Snodgrass, 2000; Peynircioglu, 1989; Pring, Freestone & Katan, 1990; Wills, Soraci, Chechile & Taylor, 2000). Peynircioglu (1989) was the first to report an SGE for pictures in a series of four experiments using normal, healthy college students. In each experiment, the SGE Application of the Self-Generation Effect 4 was explored under different generation and non-generation conditions, comparing recall for generated and non-generated picture stimuli. The picture generation conditions included drawing pictures by copying a sample picture, drawing a picture using simple written instructions and drawing pictures using connect-the-dot illustrations. Superior memory for the generated pictures was observed in all four experiments. Pring, Freestone & Katan (1990) also evaluated the role of SGE with picture stimuli. The participants were blind and sighted children. In this study, the participants recalled different pictures with raised shapes after exposure. In the self-generation condition, participants touched the raised-shape picture then proceeded to generate the picture’s name; in the non-generated condition, the labels for the raised-shape pictures were provided. Recall of the generated labels was again superior to recall of the non-generated labels. Similarly, Kinjo & Snodgrass (2000) also found a generation advantage for picture stimuli when testing for the SGE in normal adults. In this study, the naming of fragmented, incompletely drawn pictures established the generation condition, and the naming of complete pictures established the non-generation condition. Wills et al. (2000) also showed a recall advantage for pictures generated using connect-the-dot illustrations compared with traced or visually scanned pictures. However, under some circumstances, the SGE may not emerge in typically developing children and adults. Previous studies have shown that it failed to emerge with stimuli that were new, unfamiliar or meaningless (Gardiner & Hampton, 1985; Lutz, Briggs, & Cain, 2003; McElroy & Slamecka, 1982; Nairne, Pusen and Widner, 1985; Nairne & Widner, 1987; Payne, Neely & Burns, 1986). Lutz et al. (2003) found a stronger SGE for familiar clichés when contrasted with unfamiliar sentences from textbooks. When studying the SGE in monkeys using photographs as stimuli to contrast, Kornell & Terrace (2007) were only able to elicit the SGE Application of the Self-Generation Effect 5 after two weeks of generation training. The subjects were presented with a set of five novel photographs on a touch screen and were expected to touch the photographs in a pre-determined sequence. In the non-generation condition, the subjects received cues or hints about how to perform the touch sequence. In the self-generation condition, the subjects performed the sequence using only trial and error. The subjects’ performance levels were better using the non-generation condition during the first three days of training; however, as training continued, the selfgeneration condition started to show higher performance levels. The researchers concluded that “although the active generation of answers during training may result in low initial performance, it enhances long-term retention and transfer” (Kornell & Terrace, 2007, p. 685). Despite circumstances when the SGE may not emerge, the application of the SGE to persons presenting with severe aphasia still appears promising due to its successful application in people with other types of acquired brain damage (Barrett, Crucian, Schwartz, & Heilman, 2000; Chiaravalloti & DeLuca, 2002; Chiaravalloti, DeLuca, Moore & Ricker, 2005; Dick & Kean, 1989; Goverover, Chiaravalloti, Johnston & DeLuca, 2005; Lengenfelder, Chiaravalloti & DeLuca, 2003; Lipinska, Backman, Mantyla & Viitanen, 1994; Michell, Hunt & Schmitt, 1986; Multhaup & Balota 1997; O’Brien, Chiaravalloti, Arango-Lasprilla, Lengenfelder & DeLuca, 2007; Souliez, Pasquier, Lebert, Leconte & Petit, 1996). Multhaup & Balota (1997) studied the SGE with three groups, healthy elderly adults, participants with mild Alzheimer-type dementia and participants with very mild Alzheimer-type dementia, using complete and incomplete sentences as stimuli. The generation condition was established by asking the participant to complete a sentence by generating a missing word. In the non-generation condition, the examiner read a sentence to the participant. To test for the SGE, researchers used a forced-choice recognition test and a source monitoring test (Kinjo & Snodgrass, 2000), which required the Application of the Self-Generation Effect 6 participant to judge whether the sentences presented were self-generated or examiner-generated. All three participant groups demonstrated superior recall for the generated sentences. Similar results were replicated with people with mild to moderate dementia by Souliez et al. (1996), Lipenska et al. (1994) and Barrett et al. (2000). Chiaravalloti & DeLucca (2002) studied self-generation as a means of maximizing learning in multiple sclerosis (MS), a condition where memory appears to be the most consistently identified cognitive dysfunction. Because people with MS show a deficit in acquiring new information, it was expected that the SGE could provide a way for them to generate their own to-be-remembered stimuli. The participants were required to recall and recognize a list of words presented in complete and incomplete written sentences. In the nongeneration condition, the to-be-recalled word was underlined in the written sentence. In the selfgeneration condition, incomplete sentences were used; the to-be-recalled word was omitted, and participants were required to generate the missing word to complete the sentence. The results showed superior recall and recognition for the generated words in both the control group and the MS group. O’Brien et al. (2007) investigated the SGE persons presenting with traumatic brain injury (TBI) and multiple sclerosis. The researchers also aimed to describe the benefits of the SGE in participants with different types of cognitive impairment. Participants showed a strong SGE for words in both the TBI and MS groups, confirming that individuals with deficits in multiple cognitive domains after brain injury may also benefit from the SGE. Dick & Kean (1989) also investigated the SGE in patients with mild to moderate Alzheimer-type dementia and a control group of adults without dementia, using sentences as stimuli. In the generated condition, participants generated the last word in the sentence. Some Application of the Self-Generation Effect 7 letters of the word were provided to constrain responses. In the non-generated condition, the last word was presented in a larger font, requiring the participant only to read. The stimuli presentation was repeated three times. The test procedures included a free recall test and a source monitoring test including word stem completion. The results showed an SGE for the healthy control group, but no SGE for the participants with dementia. Mitchell, Hunt & Schmitt (1986) found similar results, arguing that because semantic memory is disrupted in Alzheimer’s typedementia, they did not expect to find an SGE in these participants. Such results provided support for a semantic activation theory for the SGE. Hence, this study investigated whether the SGE could be used to enhance the recognition and retention of Blissymbols in persons presenting with severe aphasia. The method for eliciting the SGE in the present study was guided by the prominent semantic activation theory, which states that the effortful act of generating a stimulus item, such as a word or a picture, activates its location in the lexical/semantic network and enhances its retrieval from memory (Graf, 1980; McElroy & Slamecka, 1982; Nairne et al., 1985; Payne et al., 1986). Early studies by Jacoby (1978) and Slamecka & Graf (1978) showed the SGE to be a function of the individual’s processing of stimuli in an elaborate way in the generation condition and in an unelaborated, non-distinctive manner in the non-generation condition. In the present study, the generation condition involved completing connect-the-dot illustrations. The to-be-recognized Blissymbols were converted into connect-the dot illustrations, as used by Peynircioglu (1989) and Wills et al. (2000). Participants generated the to-be-remembered Blissymbols by connecting the dots. In contrast, the non-generation condition involved a paired-association activity in which the participant was visually exposed to the completed Blissymbol, and its referent was named verbally. Hence, the participant’s generation of the connect-the-dot Blissymbol illustrations Application of the Self-Generation Effect 8 involved more cognitive effort than the non-generation condition. Completing the dot illustrations led to the active encoding and effortful processing of the Blissymbols at the time of generation, eliciting the SGE memory advantages during recognition and retention testing. As Kornell & Terrace (2007) and Lutz et al. (2003) suggested, the SGE may emerge over time when new, meaningless or unfamiliar stimuli (such as the Blissymbols used in this study) are used. In the current study, training and testing for the emergence of the SGE occurred across three different experimental sessions spread over a period of time, with two withdrawal periods. The first withdrawal period of one day followed experimental session one and the second withdrawal period of seven days followed experimental session two. Recognition and retention probes were used as two types of testing probes. The recognition probes occurred directly after training, thereby identifying the possible immediate recognition benefits of the SGE. The retention probes were administered after the withdrawal periods. These probe measures permitted probing for the emergence of the SGE directly after training and also assessed the influence of the SGE on the retention of previously recognized symbols over time.