Susceptibility to Memory Interference Effects Following Frontal Lobe Damage: Findings from Tests of Paired-Associate Learning
Total Page:16
File Type:pdf, Size:1020Kb
Susceptibility to Memory Interference Effects following Frontal Lobe Damage: Findings from Tests of Paired-Associate Learning Arthur P. Shimamura, Paul J. Jurica, Jennifer A. Mange&, Downloaded from http://mitprc.silverchair.com/jocn/article-pdf/7/2/144/1755204/jocn.1995.7.2.144.pdf by guest on 18 May 2021 and FeUck B. Gershberg University of California, Berkeley Robert T. Knight University of California, Davis Abstract Patients with frontal lobe lesions were adminstered tests of learning as a result of interference effects. In particular,patients paired-associatelearning in which cue and response words are exhibited the poorest performance during the initial trial of manipulated to increase interference across two study lists. In the second list, a trial in which interference effects from the one test of paired-associate learning (AB-AC test), cue words first list would be most apparent. These findings suggest that used in one list are repeated in a second list but are associated the on-line control of irrelevant or competing memory associa- with different response words (e.g., LION-HUNTER, LION- tions is disrupted following frontal lobe lesions.This disruption CIRCUS). In another test (AB-ABr test), words used in one list may be indicative of an impaired gating or filtering mechanism are repeated in a second list but are rearranged to form new that affects not only memory function but other cognitive pairs. Compared to control subjects,patients with frontal lobe function as well. lesions exhibited disproportionate impairment of second-list INTRODUCTION & Benson, 1986). Deficits in these aspects of cognition can be observed on tests of card sorting (Milner, 1964), Characterizing the underlying cognitive function associ- cognitive skill learning (Shallice & Evans, 1978), and ated with the frontal lobes has been a rather elusive motor programming (Luria, 1966). Also, deficits in the venture. Teuber, (1964) portrayed the problem as a “rid- temporal organization of information have been attrib- dle,” because neither sensory nor motoric functions by uted to frontal lobe lesions. Specifically, patients with themselves could fully describe the contribution of the frontal lobe lesions exhibit impairment in memory for frontal lobes to behavior. More recently, cognitive the temporal order in which stimuli are presented (Mil- neuroscientists have come to realize that the frontal ner, Corsi, & Leonard, 1991; Shimamura, Janowsky, lobes-more specifically, the prefrontal cortices-con- & Squire, 1990). Recentky, an assortment of memory tribute to a wide array of cognitive functions, including deficits has been observed in patients with frontal lobe aspects of attention, language, memory, and problem lesions. These deficits include problems in free recall solving (for review, see Milner, Petrides, & Smith, 1985; (Incisa della Rocchetta, 1986; Gershberg & Shimamura, Moscovitch, 1992; Shallice, 1982; Shimamura,Janowsky, in press; Janowsky, Shimamura, Kritchevsky, & Squire, & Squire, 1990;Stuss & Benson, 1986). This fact may not 1989; Jetter, Poser, Freeman, & Markowitsch, 1986), be very surprising, because the prefrontal region com- metamemory (Janowsky, Shimamura, & Squire, 1989a; prises about 28%of the cortical mantle in humans and Shallice & Evans, 1978), short-term memory uanowsky is intricately connected to many other neocortical as et al., 1989), and source memory (Janowsky, Shimamura, well as subcortical regions (Fuster, 1989;Goldman-Rakic, & Squire, 1989b). 1987). Whereas some have viewed the manifold array of Classically, prefrontal function has been associated disorders as an indication that the prefrontal cortex with planning, problem solving, and response inhibition cannot be characterized by a single cognitive process or (for historical review, see Benton, 1991;Luria, 1966;Stuss mechanism, others have emphasized the importance of 0 1995 Massachusetts Institute of Technology Journal of Cognitive Neuroscience 7: 2, pp. 144-I52 Downloaded from http://www.mitpressjournals.org/doi/pdf/10.1162/jocn.1995.7.2.144 by guest on 02 October 2021 this region in working memory or executive control. performance should occur on the first learning trial of Specifically, prefrontal cortex is thought to contribute the second list, as this trial should be most susceptible significantly to the on-line encoding, manipulation, and to interference from the first list. Finally, to determine if organization of information processing (Baddeley, 1986). interference effects were due to problems in list dis One function of working memory appears to be the crimination, subjects were given a final cued recall test active maintenance of memory representations (i.e., in which each cue word was presented and both of the short-term memory) (Fuster, 1989;Goldman-Rakic, 1987; response words were requested. If problems were sim- Jonides, Smith, Koeppe, Awh, Monoshima, & Mintun, ply due to a failure to remember which response words 1993). Another aspect appears to be the gating of extra- were associated with the first or second list, then sub- Downloaded from http://mitprc.silverchair.com/jocn/article-pdf/7/2/144/1755204/jocn.1995.7.2.144.pdf by guest on 18 May 2021 neous or irrelevant information processing (see Shima- jects should be able to report both target words when mura, 1994). For example, frontal lobe lesions appear to list information is not required. This final cued recall test affect the ability to gate or inhibit irrelevant stimulus has been used in previous studies of paired-associate information (Knight,Scabini,& Woods, 1989;Luria, 1966). learning and has been called a modified-modified free Findings of greater interference effects in the Stroop test recall (MMFR) test (Barnes & Underwood, 1959). by patients with frontal lobe lesions (Perret, 1974) also suggest that the frontal lobes may be involved in sup Results and Discussion pressing irrelevant information. Increased susceptibilityto interference may also occur Figure 2 displays paired-associate learning by patients in tasks that require the manipulation or retrieval of with frontal lobe lesions and control subjects. The data information in long-term memory. That is, patients with were subjected to a 2 x 2 x 3 ANOVA, with subject group frontal lobe lesions may exhibit increased interference (frontals and controls),list (list 1 and list 2), and learning from extraneous memory associations. To study such trial as variables. Significant main effects were observed interference effects,we analyzed performance on paired- for all three variables: subject group [F(1,16) = 12.6, associate learning paradigms commonly used in studies MSe = 44.5,p < 0.011,list [F(l,l6]) = 14.3,MSe = 17.8, of verbal learning. In the tradition of verbal learning, p < 0.011, and trial [F(2,32) = 31.5, MSe = 41.4, p < numerous test paradigms and detailed theoretical inter- 0.011. There was a significant subject x list interaction, pretations have been developed in the service of under- which occurred as a result of the disproportionately standing interference effects (for review, see Bjork, 1992; poor performance on the second list by patients with Crowder, 1976; Postman, 1961; Postman & Underwood, frontal lobe lesions [F(1,16) = 10.8, MSe = 13.5,p < 1973).In the present investigation, we assessed memory 0.011. No other effects were statistically significant Cp > interference by studying the degree to which prior learn- 0.05). It was hypothesized that a particular impairment ing affects later learning (i.e., proactive interference).We would occur on the first trial of the second list, because used two simplified versions of proactive interference it is that trial that should be most affected by interfer- paradigms. In Experiment 1, we assessed AB-AC paired- ence from memory associations acquired during first-list associate learning. In this test, cue words used in one list learning. Indeed, a signtficant subject x list interaction are used in a second list but are associated with different was observed in this analysis [F(1,16) = 6.8, MSe = 6.7, response words in a second list (e.g., LION-HUNTER, p < 0.011. LION-CIRCUS). In Experiment 2, we assessed AB-ABr Patients with frontal lobe lesions also exhibited mem- paired-associate learning, in which cue and response ory impairment on the final test of recall (MMFR test) in words used in one list are rearranged to form new pairs which cue words were presented again and both re- in a second list. sponse words were requested (see Table 2). A 2 x 2 ANOVA with subject group and list as variables revealed a significant difference between patients and control EXPERIMENT 1 subjects [F(l,l6) = 6.3,MSe = 24.5,p < 0.05l.There was Interference effects in AB-AC paired-associate learning also a significant effect of list [F(1,16) = 6.9,MSe = 9.4, were assessed in patients with unilateral frontal lobe p < 0.011, which indicated that recall of list 1 target lesions (see Fig. 1 and Table 1) and control subjects. words was poorer than recall of list 2 words. The group 'helve moderately related word pairs (e.g., RIVER- x list interaction was not significant [F(1,16) = 3.3, POND) were presented for three consecutive study-test MSe = 4.5,p = 0.081, though ceiling effects prevent learning trials. Following these trials, an additional three strong conclusions in this analysis. The results from this learning trials were administered for word pairs with the final recall test suggest that the impairment in paired-as- same cue words but with different response words (eg., sociate learning observed in patients with frontal lobe RIVER-BROOK). To the extent that the prefrontal cortex lesions extends beyond a problem of list discrimination. acts to gate or filter interference from prior memory For example, if the problem was simply due to a failure associations, patients with frontal lobe lesion should to determine if a remembered response word came from exhibit disproportionate impairment on second-list the first list or the second list, the patients may have learning. Moreover, the most significant decrement in been able to recall as many words as control subjects Shimamura et al.