Brain, Vol. 126, No. 5, 1112-1126,
May 2003
© 2003 Guarantors of Brain
doi: 10.1093/brain/awg112
Source versus content memory in patients with a unilateral frontal cortex or a temporal lobe excision
1 Victoria General Hospital, Victoria, British Columbia and 2 Montreal Neurological Institute and Department of Psychology, McGill University, Montreal, Quebec, Canada
Correspondence to: Laila Thaiss, Department of Pediatric Psychology, Victoria General Hospital, 1 Hospital Way, Victoria, British Columbia V8Z 6R5, Canada E-mail: laila.thaiss{at}mail.mcgill.ca
Received June 7, 2002. Revised December 3, 2002. Accepted December 27, 2002.
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
It has been suggested previously that patients with a frontal lobe lesion might have a specific impairment in the retrieval of the source of information despite adequate memory for facts. Patients with an anterior temporal excision are known to have impairments in memory for facts and the question arises as to whether they are also impaired in source memory. The present study compared memory for facts and their source in patients with a unilateral frontal cortical or an anterior temporal excision in a situation in which both types of information were encoded explicitly. Patients with a unilateral frontal cortex or a temporal lobe excision watched videos of a game show and were instructed to attend to both the trivia facts and their source (the identity of the speaker or the relative time of presentation). Patients with a frontal cortex excision were not impaired on either fact or source memory. This was true even when a subgroup of patients with an excision involving the dorsolateral frontal cortex was examined. In contrast, patients with a left temporal lobe excision were impaired in both fact and identity source memory and right temporal lobe patients were impaired in identity source memory. All patients performed similarly to normal controls in temporal source memory. The present results are consistent with the view that source information is part of an associative network of information about an episode and that the medial temporal region is critical for both source and content memory. Furthermore, if source information is encoded explicitly, the frontal cortex does not appear to be necessary for its retrieval. Instead, it is proposed that the frontal cortex plays a metacognitive role in memory retrieval.
Keywords: source memory; frontal cortex; anterior temporal lobe; strategic processes; temporal context
Abbreviations: DLFC = dorsolateral frontal cortex; LF = left frontal cortex; LT = left temporal lobe; RAVLT Rey Auditory Verbal Learning Test; RF = right frontal cortex; RT = right temporal lobe
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
One of the defining features of episodic memory is the association between one’s knowledge or memory for the content of an event and the memory for the source or context in which it was acquired (Tulving, 1983
, 1985). It has been proposed by some researchers that source memory may be dissociable, in terms of its functional localization, from content or fact memory (Schacter et al., 1984; Shimamura and Squire, 1987, 1991; Janowsky et al., 1989; Glisky et al., 1995). Furthermore, it has been argued that impairments of source memory, in the absence of a severe impairment of fact memory, are characteristic of patients with a frontal lobe lesion and of elderly normal subjects who perform poorly on tests of frontal lobe function (Janowsky et al., 1989; Craik et al., 1990; Shimamura et al., 1991; Glisky et al., 1995; Schacter, 1995). Various studies examining the ability to remember temporal order or to judge the relative recency of events in memory have also shown deficits in patients with frontal cortex lesions (e.g. Milner, 1971; Làdavas et al., 1979; Shimamura et al., 1990; Kesner et al., 1994; Kopelman et al., 1997). These studies have frequently been cited as further evidence of the specificity of the frontal cortex to source memory (e.g. Schacter, 1995; Kopelman et al., 1997; Schacter et al., 1998).
It is known, however, that the medial part of the temporal lobe also plays a critical role in episodic memory (Milner, 1958; Squire et al., 1993; Gabrieli, 1998). The hippocampus, in particular, is important for associating the content of an event with its source or context (Eichenbaum et al., 1994; Mishkin et al., 1997; Vargha-Khadem et al., 1997; Murray, 2000). The hypothesis that both source and content memory are processed by the medial temporal lobe system is supported by studies showing impairments in both types of memory in patients with a temporal lobe lesion (Kopelman et al., 1997; Schwerdt and Dopkins, 2001). Furthermore, many studies have found strong correlations between performance on tasks of source and content memory in both patients and normal adults (Shimamura and Squire, 1991; Cycowicz et al., 2001; Schwerdt and Dopkins, 2001). Some studies of amnesic patients have also found that memory for context is more closely associated with the severity of the amnesia than with frontal lobe function (Kopelman, 1989; Pickering et al., 1989)
These apparently conflicting findings regarding the neural substrate of source memory are best explained in a cognitive model proposed by Johnson and colleagues (1993). These authors have proposed that source information is part of an associative network of characteristics about an episode. Depending on the quality of the encoded information, attribution of source is proposed to require varying degrees of judgement and strategic search of this network during retrieval (Johnson et al., 1993). Patients with a frontal cortical lesion are known to be impaired in using strategic processes to aid retrieval (Petrides and Milner, 1982; Jetter et al., 1986; Hirst and Volpe, 1988; Incisa della Rocchetta and Milner, 1993; Gershberg and Shimamura, 1995; Mangels, 1997). It is therefore possible that the deficits in source memory reported in such patients were secondary to the strategic processing demands of the memory tests. It should be noted that, in several studies of patients with a frontal lobe lesion, the processing demands for content and source retrieval were different because the subjects were explicitly asked to learn new facts (content information) but were unaware that the source of the information was relevant (e.g. Janowsky et al., 1989; Shimamura and Squire, 1991). In other studies involving the judgement of temporal recency, the task requirements placed high demands on strategic relational processing. Specifically, these tasks required participants to keep track of temporal relationships between a long list of items, but the individual items were not associated with any distinct contextual cues (see also McAndrews and Milner, 1991). Thus, it remains to be determined whether patients with frontal cortical damage would be impaired on source but not fact memory if the strategic organizational or relational demands were reduced in both the source and content memory tasks.
In addition to the issue of the processing demands in the source and context memory tasks used to date, there have been very few studies that have directly compared the performance on such tasks of patients who have a frontal cortex excision with those who have an anterior temporal lobe excision. In the few studies that exist (e.g. Kopelman et al., 1997; Melo et al., 1999), many of the patients with a frontal lobe lesion had additional damage to regions outside the frontal cortex, such as the basal forebrain or diencephalon, which are closely associated with the medial temporal lobe memory system. The fact that there was significant damage to non-frontal structures in the frontal lobe groups in these studies renders the interpretation of the results difficult with respect to the specific contribution of the frontal cortex per se to source memory and to the accurate retrieval of events.
In the present study, we compared patients with a unilateral excision restricted to the frontal cortex or the anterior temporal lobe, and normal control subjects. All participants watched two different videos of a trivia game show and were immediately tested on their memory for the answers to the trivia questions and for the source of these answers. Two types of source information were tested in separate conditions: information about who provided the fact (identity video condition) and information about when the fact was provided (temporal video condition). In order to emphasize the importance of both source and fact information for the task, subjects were specifically instructed to attend to the trivia fact (‘what was the answer to the question?’) and its associated source (‘who gave it?’ or ‘when was it given?’) while watching the video. It was assumed that directing attention to both types of information, and thus treating both as equally important, would make it more likely that the quality of encoding of the two types of information would be similar (Johnson and Raye, 1981; Lindsay and Johnson, 1989). Furthermore, because explicit, associative encoding of the source and content information was encouraged, subjects were expected to be less reliant on strategic processes at retrieval (e.g. Johnson et al., 1993).
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
The patients in this study had undergone a unilateral cortical excision for the relief of pharmacologically intractable focal epilepsy. There were 18 patients with an excision restricted to the prefrontal cortex. Eight of these patients had a left frontal (LF) and 10 had a right frontal (RF) cortex excision. There were also 30 patients with an excision from the anterior temporal lobe: 16 with a left-sided (LT) and 14 with a right-sided (RT) excision. The control group comprised 20 neurologically normal individuals, many of whom were friends or family members of the patients studied. The subjects gave informed consent according to the Declaration of Helsinki for participation in this study and for the release of MRI and neuropsychological assessment results. The study was approved by the ethics committees of the Montreal Neurological Institute and the Department of Psychology, McGill University.
All subjects were fluent speakers of French or English. Bilingual subjects were tested in the language in which they had conducted most of their formal education. The gender, age, language, education and full-scale IQ of the subject groups are provided in Table 1. Full-scale IQ scores and neuropsychology test results were available only for the four patient groups. The groups did not differ significantly with respect to age [F(4,63) = 1.54, not significant (n.s.)], length of education [F(4,63) = 0.38, n.s.] or full-scale IQ [F(3,44) = 0.30, n.s.]. The Rey Auditory Verbal Learning Test (RAVLT) is a standardized test of verbal memory in which a patient must learn a list of unrelated words over five trials, then recall the list after an interference list and again after a 30 min delay. Results of this test were available for 39 of the 48 patients (Table 2). Those patients who were not tested on the RAVLT had been tested with the Wechsler Memory Scale, Logical Memory Subtest (WMS-LM), a standardized memory test in which subjects learned a short story in one trial and recalled it after a 30 min delay. A one-way ANOVA (analysis of variance) examining the results of the delayed memory trial on the RAVLT showed that LT patients were significantly impaired compared with LF, RF or RT patients [F(3,38) = 14.52, P < 0.001]. See Table 2 for mean scores and standard deviations on the RAVLT for each group. The mean scores on the RAVLT of the LT patients, but not those of the other three patient groups, were considered to be significantly impaired compared with the norms for the test. Similarly, of the nine patients who received the WMS-LM rather than the RAVLT as part of their neuropsychology test battery, only the LT patients had scores that were significantly impaired compared with the norms of this test.
|
|
Inclusion of subjects in this study was strict in order to ensure that the results obtained could be attributed with confidence to damage of the prefrontal or anterior temporal region. Patients were excluded from the study if they met one or more of the following criteria: (i) seizure disorder arising from or associated with diffuse cerebral damage; (ii) bilateral independent electrographic abnormalities; (iii) rapidly growing neoplasm; (iv) bilateral or right hemisphere speech representation, determined by the preoperative intracarotid amobarbital procedure; (v) damage contralateral to the side of operation and/or damage involving subcortical structures (e.g. basal ganglia, basal forebrain region); (vi) full-scale IQ <80 on the WAIS-R (Wechsler Adult Intelligence Scale—Revised); (vi) age >55 years; (viii) impaired receptive or expressive language abilities.
Frontal cortex excisions
A postoperative MRI of each patient’s brain was available for 14 of the 18 patients with an excision from the prefrontal cortex. For those patients without a postoperative MRI, the location and extent of the excision was determined from a preoperative MRI (one subject) or from operation reports and drawings of the excisions provided by the neurosurgeon (three subjects). The available MRIs were transformed into the Talairach standardized proportional stereotaxic space (Talairach and Tournoux, 1988) in order to compare the size and location of the excisions across patients. The location and extent of the surgical resections within the frontal cortex were identified and labelled using DISPLAY, an interactive 3D drawing program developed at the Montreal Neurological Institute (for details see Koski et al., 1998). Briefly, a mouse-driven screen paintbrush was used to highlight those voxels in the MRI volume that corresponded to the resected region as determined by visual inspection. The internal boundary of the excision (i.e. the distinction between excision and adjacent remaining brain tissue) was defined by the voxels whose MRI intensity fell within the range of values representing CSF using a histogram procedure. Each excision was labelled on the coronal sections in steps of 1 or 5.5 mm (when high-resolution images were unavailable), beginning at the frontal pole and ending at the caudal extent of the excision. The external limit of each excision (i.e. excision versus CSF/skull) was determined by applying a mask to the subjects’ MRI, which represented the contour of an average brain based on 305 MRIs of healthy young volunteers (Evans et al., 1996). The locations and extents of the excisions from the left or right frontal cortex are illustrated in Figs 1 and 2, respectively. For the three patients for whom an MRI of the brain was not available, the location and extent of the excision was illustrated in drawings made by the neurosurgeons at the time of surgery (patients H.S., C.C. and B.G.).
|
|
The aetiology of the seizures included low-grade tumour (n = 5), arteriovenous malformation (n = 1), cavernous haemangioma (n = 4) and idiopathic epilepsy (n = 8). The anterior speech area was spared in all patients with a left frontal excision. The mean time elapsed since surgery for the patients in the frontal group was 5 years and 8 months, with a range of 3 months to 25 years.
Temporal lobe excisions
The unilateral surgical removals in this group of patients involved either a selective amygdalohippocampectomy (seven patients with a left- and four with a right-sided excision) or an excision of the anterior temporal cortex that included the amygdala and hippocampus (nine left- and 10 right-sided excisions). In the selective amygdalohippocampectomy procedure, a narrow excision was made through the first or second temporal gyrus or through the floor of the superior temporal sulcus, allowing the surgeon to remove the amygdala (four-fifths or total resection) and the anterior 2.5–3.5 cm of the hippocampal formation together with the surrounding parahippocampal gyrus (2.5–4 cm), while sparing the rest of the temporal neocortex (Olivier, 1988). For the anterior temporal lobe removals, the excisions of temporal cortex ranged from 4 to 6 cm along the sylvian fissure and from 4.5 to 6 cm along the base of the temporal lobe. In these excisions, the amygdala was also removed, as was the anterior portion of the hippocampus (between 1.5 and 4 cm) and parahippocampal cortex (between 1.5 and 4 cm). Thus, in both the anterior temporal lobe group and the group with selective amygdalohippocampal excisions a significant proportion of the amygdala, hippocampus and adjacent parahippocampal cortex was removed. It was therefore not possible to examine the data for any differences in the contribution of these medial temporal limbic neural structures to the types of memory examined here. Figure 3 shows an example MRI (in standardized proportional stereotaxic space) of each of the two types of anterior temporal lobe excisions represented in this group of subjects.
|
The aetiology of seizures in this group of patients included idiopathic epilepsy (n = 13), febrile illness (n = 4), head trauma (n = 5), low-grade tumour (n = 5), perinatal illness (n = 2) and meningitis (n = 1). The mean time elapsed since surgery for the patients with a temporal lobe excision was 6 years and 9 months, with a range of 3 months to 30 years.
Materials and procedures
All subjects participated in two testing conditions. In both conditions they watched a video of a trivia game show and were then tested for their memory of the events in the videos. In the ‘identity’ condition, the video showed three contestants taking turns to answer obscure trivia questions posed by the host of the game. In the ‘temporal’ condition, the video showed a man answering trivia questions posed by the host of the game in three temporally distinct sections. Subjects were asked to attend to the answer provided, as well as to who answered the question (identity condition) or when the question was answered (temporal condition). Immediately following the presentation of each video, subjects were tested on cued recall of the answers to the trivia questions and on the source of the information (i.e. who answered the question or when it was answered). In addition, forced-choice recognition of the trivia facts was tested for those questions that the subject had failed to answer correctly in the cued recall test.
The material for this study was developed both in English and in French. In each language version, attention was paid to linguistic or cultural factors that might have made the answer to a trivia question more familiar or memorable. Two sets of 36 obscure trivia questions were developed, each with an equivalent number of questions, on the topics of music, sports, geography and science. The questions were taken from the advanced version of the game Trivial Pursuit®, from text books and from the internet. The questions for the study were chosen after first testing a larger sample of trivia questions on a group of 15 graduate students and postdoctoral fellows. The questions that these subjects were unable to answer were used in the study. A small pilot study involving undergraduate students also helped to determine the difficulty level of the questions and final adjustments were made following this pilot study. Two versions of each video condition (identity and temporal conditions) were recorded for the two sets of test questions, using the same actors in each version. This allowed the set of trivia questions used in each video condition to be counterbalanced across subjects.
In the identity condition, the videotape began with three contestants being introduced by name to the ‘audience’ by the host of the game, who was a professional-looking young woman wearing a casual suit. The contestants were a young, extroverted woman dressed in brightly coloured clothes, a middle-aged conservative, awkward gentleman in a suit and brightly coloured tie, and an enthusiastic teenage girl in a baseball cap and T-shirt. The host then began the trivia contest and proceeded to ask the questions in random order. Each time a question was asked, one of the contestants would press a buzzer in order to be identified by name by the host and would then answer the question. The order in which contestants answered the questions in the video was random, with the restriction that they did not answer more than two questions in a row. Unlike a true game show, the contestants always answered the question correctly and, by the end of the show, each had answered an equal number of questions from each of the four topics. The videotape lasted 
8 min.
In the temporal condition, the videotape began with the host introducing a young man by name and explaining that he was practising for the world trivia championship. As in the previous condition, the host then began the practice and proceeded to ask the trivia questions in random order. The young man would press a buzzer before answering the question and would always answer the question correctly. In this condition, the 36 questions were answered in three successive parts, each containing an equal number of questions from each of the four topics. The host announced the end of each part and the beginning of the next with the statements: ‘This concludes the [first/second/ third] part of our trivia practice’ and ‘Welcome to part [two/three] of our trivia practice. Pete, are you ready to begin?’. In the intervals between the three parts the video screen was blank for 5 s. This videotape also lasted 8 min. Different actors played the parts of the contestants and the hosts in the two video conditions and in the two language versions. Although the two video conditions were matched in terms of the general structure of the game show, it is important to note that there was more perceptual information available for distinguishing between and encoding the identity of the speakers in the identity condition than for distinguishing between the times (first, second or third part) in the temporal condition. This is due to the inherently abstract nature of temporal information. It was expected, therefore, that source recognition would be somewhat more difficult in the temporal condition than in the identity condition.
While viewing the videos, the subjects had a buzzer available to them and were instructed to press it if they happened to know the answer to a trivia question. On the rare occasions that the subject knew the answer and pressed the buzzer, the video was immediately stopped in order to give the subject time to answer the question. If the subject answered the question correctly, this item was dropped from his/her subsequent fact recall and fact recognition tests. Source recognition was, however, tested on these items. Immediately following the presentation of each video, subjects were read each of the questions that had been asked of the contestants (in a different order from that of the video presentation). For each question, they were asked to recall the answer to that question. Regardless of whether the subjects correctly recalled the answers to the trivia questions, they were then asked about the source of the answer: ‘who answered that question’ (in the identity condition) or ‘when was that question answered’ (in the temporal condition). Pictures of the three contestants with their names written at the bottom of the picture, in the identity condition, or the numbers ‘1’, ‘2’ and ‘3’ for the three parts, in the temporal condition, were laid out on cards during the test. After the subjects had been tested on fact recall and source recognition for the trivia questions, a three-choice recognition test of the trivia facts was given for each of the questions that the subject had failed to answer correctly in the fact recall test. The three choices in this test included the correct answer and two distractor items that were plausible alternatives to the correct answer. Trivia facts that had been correctly recalled were scored as correct in the recognition test. Responses to the trivia questions (both fact recall and recognition) were calculated as the proportion correct out of the total number of questions asked. Responses in the source recognition test were scored independently of those for fact recall or recognition and were also calculated as the proportion correct.
Language versions
A comparison of the English and French versions of the test material was made using 26 normal control subjects (nEnglish = 14, nFrench = 12). The difficulty of the questions in the two language versions of the tests was examined using independent-sample t tests. The dependent variables tested were the mean fact recall scores and the fact recognition scores (with the answers to the questions from the two conditions combined), and the two source recognition scores (for ‘who’ and ‘when’). There were no significant differences in memory for the trivia facts for the English and French versions of the test [recall, t(24) = 1.63, P = 0.12; recognition, t(24) = 1.68, P = 0.11]. There were also no significant differences between the English and French language versions in subjects’ ability to remember the source of the information [‘who’, t(24) = 0.51, P = 0.61; ‘when’, t(24) = 0.07, P = 0.94].
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In order to address the main question of the study, performance on the source recognition test was compared with that of the fact recognition test since both were three-item forced-choice tests. Fact recall was analysed separately. Individual subject scores for patients with a frontal cortical excision on each of the tests are presented in Table 3. All post hoc testing described in the following section used the Newman–Keuls method, with a significance level of 0.05. Simple effect analyses used the pooled error term and degrees of freedom.
|
Subjects’ general knowledge
On average, the subjects in the five groups knew the answers to fewer than one of the 36 trivia facts (0.74/36 answers, averaged over the two conditions) before the answer was provided by the contestant on the video. There were no significant group differences in the subjects’ prior knowledge of the trivia facts for either the identity condition [F(4,63) = 0.91, P = 0.46] or the temporal condition [F(4,63) = 1.02, P = 0.4], as tested using one-way ANOVAs.
Cued recall of trivia facts
Differences between the groups in the proportion of trivia facts that were correctly recalled were examined using a two-way repeated measures ANOVA, in which group was the between-subjects factor and the video condition (identity and temporal conditions) was the within-subjects factor. This analysis revealed a significant main effect of group [F(4,63) = 5.88, P < 0.001]. The main effect of condition and the interaction effect were not significant. Post hoc testing showed that LT patients were impaired in recalling the answers to the trivia questions compared with all other subject groups. The LF, RF and NC subjects did not differ significantly from one another (Fig. 4).
|
Comparison of source and fact recognition
A three-way repeated measures ANOVA was used to examine the proportion of correct recognition responses across the five groups of subjects in each of the two video conditions (identity and temporal) and the two tests (fact and source recognition). This analysis revealed a significant three-way interaction [F(4,63) = 2.44, P = 0.05]. The two-way interaction between the two video conditions and the recognition tests [F(1,63) = 78.2, P < 0.001] and the main effects of group [F(4,63) = 4.91, P < 0.01], condition [F(1,63) = 37.62, P < 0.001] and recognition test [F(1,63) = 536.27, P < 0.001] were also significant. The two-way interactions of group x video condition and group x recognition test were not significant.
For the two-way interaction between the video conditions and the recognition tests, post hoc testing showed that, overall, subjects were significantly better at recognizing who had answered the question (identity source) than when the question was answered (temporal source). Although the temporal source test was more difficult than the identity source test, performance on the temporal source test was significantly above the 0.33 chance level [mean 0.48; difference between mean and chance score, t(67) = 13.85, P < 0.001]. Fact recognition was also significantly easier than source recognition in both video conditions (Fig. 5).
|
Simple effect analyses were used to examine the three-way interaction. Simple interactions between the groups and the tests (source versus fact recognition) in each video condition were not significant, nor were the simple interactions between the groups and the video conditions (identity versus temporal) for each of the tests significant. Simple main effect analyses of the groups, when tested on fact recognition in the identity and the temporal video conditions, showed significant group differences [identity condition, F(4,87) = 3.75, P < 0.01; temporal condition, F(4,87) = 3.75, P < 0.01]. Post hoc testing revealed that, in both the identity and temporal conditions, LT patients were impaired in recognizing the correct answer to the trivia question compared with all other subject groups. These four subject groups (RT, LF, RF and NC) did not differ significantly from one another in their fact recognition performance in either video condition.
In the source recognition test of the identity condition, simple main effect analyses of the groups also showed a significant difference [F(4,87) = 5.0, P < 0.01]. LT subjects were significantly impaired in their recognition of who had responded to the question compared with NC, RF and LF subjects (Fig. 6). In addition, RT subjects were significantly impaired in recognizing the source (‘who’) compared with RF and NC subjects. LF and RF subjects did not differ significantly from NC subjects in their memory for identity source information. There was no significant difference between the groups in subjects’ source memory in the temporal condition (‘when’).
|
Correlations between memory for source and for fact
The correlation between source recognition and fact recognition in the identity condition and the temporal condition was examined in each of the subject groups. These correlations are listed in Table 4AA. As can be seen from this table, the correlation (r) in the identity condition was moderate to strong for each of the five groups. In the temporal condition, the correlations for the LF and RF patients and normal control subjects were also moderate. This contrasts with the very weak correlations between memory for the trivia facts and temporal source in LT and RT patients.
|
Differences in the strength of these correlations across the two video conditions were examined statistically for each patient group. Since the pattern of these correlations in the identity and temporal conditions was similar for patients with excisions from the same brain region, regardless of the lateralization of their excision, the groups of patients with LT and RT excisions and those with LF and RF cortex excisions were combined. The individual subjects’ scores on the recognition tests were transformed into Z scores in order to remove the effect of the differences in mean recognition memory scores between groups of patients with left- versus right-hemisphere excisions. The correlations between source and fact recognition for each of the two video conditions were again calculated from these standardized scores for each of the combined groups. The significance level of the difference in the strength of these correlations was examined using Fisher’s Zr. Table 4BB lists the correlations and the difference between the correlations for each group. As can be seen from this table, the correlations between source and fact recognition were not significantly different for the identity and the temporal conditions in patients with frontal cortex excisions [Z = 0.03, n.s.] and normal control subjects [Z = 1.88, P = 0.07], although in the latter group the correlation in the temporal condition showed a tendency to be weaker than in the identity condition. In contrast, for the temporal lobe patients there was a significantly weaker correlation between recognition of source and fact in the temporal condition as opposed to the identity condition [r = 0.02 versus r = 0.67, respectively; Z(r1–r2) = 2.85, P < 0.001].
|
Source versus fact recognition in patients with DLFC excision
Because there was heterogeneity in the location of the excisions in the frontal cortex groups, and the frontal cortex is known to consist of multiple functionally distinct regions, we also examined memory performance in a subgroup of these patients. Specifically, source versus fact recognition in the two video conditions was examined in a more homogeneous subgroup of patients whose lesions extended into the dorsolateral frontal cortex (DLFC), involving the middle and superior frontal gyri (i.e. areas 9, 10 and 46). In Table 3, an asterisk beside a subject’s initials indicates the patients that were included in this subgroup. Since the two groups of patients with left- and right-sided DLFC excisions were too small for statistical analysis, these subjects were combined into a single group. Patients with excisions that included the DLFC were compared with those with LT and RT excisions and with normal control subjects. Differences between the groups in the proportion of trivia facts that were correctly recalled were examined using a two-way repeated measures ANOVA, in which group was the between-subjects factor and the video condition (identity and temporal conditions) was the within-subjects factor. This analysis revealed a significant main effect of group [F(3,57) = 5.65, P < 0.001]. The main effect of condition and the interaction effect were not significant. Post hoc testing showed that LT patients were impaired in recalling the answers to the trivia questions compared with all other subject groups. The DLFC, RT and NC subjects did not differ significantly from one another (Fig. 7).
|
A three-way repeated measures ANOVA was used to examine the proportion of correct recognition responses across the four groups of subjects (DLFC, LT, RT and NC) in each of the two video conditions (identity and temporal) and the two tests (fact and source recognition). The results of this analysis were very similar to those described above for the five subject groups. There was a significant three-way interaction [F(3,57) = 3.13, P < 0.05]. The two-way interactions between the two video conditions and the recognition tests [F(1,57) = 73.3, P < 0.001] and between recognition tests and groups [F(1,57) = 3.2, P < 0.05] were also significant, as were the main effects of group [F(3,57) = 5.5, P < 0.01], condition [F(1,57) = 33.7, P < 0.001] and recognition test [F(1,57) = 542.7, P < 0.001]. The two-way interaction of group x video condition was not significant.
As described previously, subjects were significantly better at recognizing who had answered the question than when the question was answered (two-way interaction between the video conditions and the recognition tests). Results of the simple main effect analysis based on the three-way interaction showed significant differences between the groups on fact recognition in the identity and in the temporal condition [F(3,57) = 2.73, P = 0.05 and F(3,57) = 3.6, P < 0.05 respectively] and between the groups on source recognition in the identity condition [F(3,57) = 3.2, P < 0.05] (Fig. 8). Specifically, for both the identity and temporal video conditions, LT patients were significantly impaired compared with all the other groups in their recognition of the answers to the trivia questions (fact recognition). RT, DLFC and NC subjects did not differ significantly from each other on this task. LT patients were also significantly impaired compared with the DLFC patients and normal control subjects in their ability to recognize who gave the answer in the identity condition. RT, DLFC and NC subjects did not differ significantly from each other on this task. Simple main effect analysis showed no significant group differences for source recognition in the temporal condition.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In the present study, trivia facts and their source (the identity of the speaker or the relative time of presentation) were encoded explicitly in order to encourage a strong memory trace for both aspects of the episode. Thus, the question could be posed whether patients with an excision restricted to the frontal cortex and patients with an anterior temporal lobe excision involving the amygdala and hippocampal/parahippocampal region would perform differentially on content versus source memory when both aspects of the information were encoded in a comparable manner. Patients with a frontal cortex excision were not impaired on either fact memory or its source. This was true even when a subgroup of the patients whose excision involved the dorsolateral prefrontal cortex was examined. In contrast, patients with a LT excision were impaired on both memory for the trivia facts and the identity of the speaker. The normal performance of the patients with a frontal excision on both source and fact memory when both types of information were encoded explicitly clarifies the contribution of the frontal cortex in memory. In earlier studies reporting poor performance on source memory, but not fact memory, in patients with a frontal lesion, the subjects were instructed to learn only the facts. Thus, memory of the source was encoded incidentally. Since the source of the fact was not a strong part of the associative network of the encoded episode, it was necessary for subjects to use active search strategies to retrieve this information. Under these circumstances, subjects with a frontal lesion may have performed poorly in the source memory conditions, because they were less likely to initiate an effective search strategy to retrieve the poorly encoded information. This view is consistent with evidence that the frontal cortex plays a major role in the strategic search for mnemonic information (e.g. Incisa della Rocchetta and Milner, 1993
; Gershberg and Shimamura, 1995; Petrides, 1996). Recent functional neuroimaging data have also shown that the prefrontal cortex is critical for disambiguating mnemonic information in situations where the memory traces cannot be clearly and unambiguously driven by external stimuli (e.g. questions) or specific contexts (Cadoret et al., 2001).
Patients with left temporal excision were impaired in recalling both the trivia facts and the identity of the speaker who provided the source of this information. Furthermore, memory for the trivia facts and the identity of the speaker were strongly correlated for all subject groups (Table 4A4A). These findings are consistent with the view that declarative memories, including those for source information, are dependent on the medial temporal region (Milner, 1972; Squire and Zola-Morgan, 1991; Gabrieli, 1998).
A dissociation between memory for facts and identity source was found in patients with a right temporal excision. These patients were impaired in remembering who had answered the question compared with those with a right frontal (RF) excision and normal control subjects, but were not impaired on fact recognition or recall (Figs 4 and 6, respectively). These results are consistent with those of a recent study by Schwerdt and Dopkins (2001), who found LT patients to be impaired in their memory for both actions and the source of these actions, whereas RT patients had difficulty remembering the source of the actions, but not the actions themselves. In the present study (and in that of Schwerdt and Dopkins, 2001) much of the information contributing to source discrimination and encoding was non-verbal (e.g. facial features, observed mannerisms and the pitch of the voice in the present study, and non-verbal features associated with discriminating the medium of object presentation in the study of Schwerdt and Dopkins). It is possible that the source memory impairment of the patients with a RT excision was the result of their known deficits in non-verbal memory (e.g. Milner, 1958; Milner, 1980).
Temporal source recognition, although above chance in the present study, was significantly more difficult than memory for identity source for all subject groups. This finding was probably due to the relative paucity of perceptual information in the temporal condition (hence poorer encoding of this information) compared with the identity condition. Specifically, whereas the perceptual experience necessary for the discrimination and encoding of source information was elaborate in the identity condition (i.e. distinct voices, facial features, mannerisms, names), in the temporal condition it was relatively impoverished, consisting primarily of the host’s announcement of the beginning and end of the individual parts of the game.
Patients with a frontal cortex excision, including those with DLFC involvement (Figs 6 and 8, Table 3), had a level of performance similar to that of control subjects on temporal source memory. Previous studies had found that such patients were impaired in judging the relative recency of stimuli (Làdavas et al., 1979; Smith and Milner, 1983; Shimamura et al., 1990; Milner et al., 1991; Kopelman et al., 1997). Unlike earlier studies, however, in the present task there were three distinct time periods that were clearly delineated. The subject was specifically instructed to form an association between the fact and the time period within which it occurred. Thus, the test was one of recalling the distinct temporal context within which an event occurred. By contrast, in the recency task that yielded reliable deficits in frontal cortex patients (e.g. Milner, 1971; Milner et al., 1991), a long list of item-pairs was presented to subjects in an arbitrary order. In this test, each item-pair occurred at a different point in time relative to the other items and there were no distinct temporal divisions with which to associate the items. McAndrews and Milner (1991) found that frontal cortex patients, who were impaired on the recency judgement task when simply asked to name the objects during encoding, performed normally when required to manipulate the objects actively at the presentation trial. These authors suggest that, in the latter situation, manipulating the object helped subjects create a temporally distinct memory trace of the event, which facilitated memory for temporal order. In a study by Mangels (1997) examining memory for temporal order in patients with a frontal lobe lesion, the structural organization of the information, rather than the context memory per se, was important in predicting the performance of these subjects.
An interesting finding in the present study was the difference in the pattern of performance of patients with a temporal lobe excision compared with frontal cortex patients and normal controls in their memories for who provided the fact, as opposed to when it was provided. Whereas frontal cortex patients were not impaired in either source memory task, those with a temporal lobe excision showed a difference in their performance on the two tasks: there was no difficulty on temporal source recognition, but an impairment in remembering the perceptually richer identity source information. A similar dissociation between temporal context memory and, in this case, spatial context memory was found in a study by Kopelman and colleagues (1997). In their study, the performance of temporal lobe patients was similar to that of controls on the former but was impaired on the latter.
Results of the correlation analyses between temporal source and fact recognition for each group sheds light on these findings. Although there was a moderate correlation between recognition of the trivia facts and their temporal source for frontal cortex patients and normal control subjects, there was no correlation between these two measures for temporal lobe patients. The strength of the correlation between source and fact recognition was significantly different in the two video conditions for the temporal lobe group, but not for the frontal cortex patients. It appears that, although all subjects were able to retrieve temporal source information to a similar degree, the method by which they did this differed. One possible explanation is that temporal lobe patients actively manipulated items and judged the relative salience of the facts, without necessarily remembering the individual items or their association with the three distinct time periods. In contrast, the significant correlation between source and fact memory in frontal cortex patients suggests that they remembered the specific association between the fact and the time period identified in the videotape. The normal control subjects showed a moderate level of correlation between their memory for the fact and the temporal source of the item, and the difference in the strength of this correlation compared with that in the identity condition was marginally significant. The way in which normal subjects solved this task, therefore, may have varied depending on the strength of the trace for the event and the ease with which individuals made associations between the question and its source.
In summary, the results of the present study are consistent with the view of an associative network of information about an episode that contains information about content and source. This information is encoded by structures in the medial temporal lobe, leading to impairments in both source and content memory in patients with lesions of these structures. Patients with a frontal cortex excision appear not to be impaired in either source or content memory if this information is encoded explicitly and unambiguously. The present results suggest that memory for source is likely to be processed by the same neural substrates that encode and retrieve content memory. Furthermore, the results are consistent with several views that attribute to the hippocampus and the parahippocampal cortex an important role for associating the content of an event with its source or context (Eichenbaum et al., 1994; Mishkin et al., 1997; Murray, 2000). The contribution of the frontal cortex to source memory reported in some previous studies may have reflected metacognitive contributions to memory retrieval (Petrides, 1996) rather than being a necessary component of source memory per se. Similarly, the activation in the prefrontal cortex reported in some neuroimaging studies of memory may have also reflected metacognitive processes that could contribute to source attribution in some situations (e.g. Henson et al., 1999; Ranganath et al., 2000).
|
Acknowledgement |
|---|
The present study was supported by the Canadian Institutes for Health Research.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Cadoret G, Pike GB, Petrides M. Selective activation of the ventrolateral prefrontal cortex in the human brain during active retrieval processing. Eur J Neurosci 2001; 14: 1164–70.[CrossRef][Web of Science][Medline]
Craik FIM, Morris LW, Morris RG, Loewen ER. Relations between source amnesia and frontal lobe functioning in older adults. Psychol Aging 1990; 5: 148–51.[CrossRef][Web of Science][Medline]
Cycowicz YM, Friedman D, Snodgrass JG, Duff M. Recognition and source memory for pictures in children and adults. Neuropsychologia 2001; 39: 255–67.[CrossRef][Web of Science][Medline]
Eichenbaum H, Otto T, Cohen NJ. Two functional components of the hippocampal memory system. Behav Brain Sci 1994; 17: 449–518.[Web of Science]
Evans AC, Collins DL, Holmes CJ. Computational approaches to quantifying human neuroanatomical variability. In: Toga AW, Mazziotta JC, editors. Brain mapping: the methods. San Diego: Academic Press; 1996. p. 343–61.
Gabrieli JDE. Cognitive neuroscience of human memory. Annu Rev Psychol 1998; 49: 87–115.[CrossRef][Web of Science][Medline]
Gershberg FB, Shimamura AP. Impaired use of organizational strategies in free recall following frontal lobe damage. Neuropsychologia 1995; 13: 1305–33.
Glisky EL, Polster MR, Routhieaux BC. Double dissociation between item and source memory. Neuropsychology 1995; 9: 229–35.[CrossRef][Web of Science]
Henson RNA, Shallice T, Dolan RJ. Right prefrontal cortex and episodic memory retrieval: a functional MRI test of the monitoring hypothesis. Brain 1999; 122: 1367–81.
Hirst W, Volpe BT. Memory strategies with brain damage. Brain Cogn 1988; 8: 379–408.[CrossRef][Web of Science][Medline]
Incisa della Rocchetta A, Milner B. Strategic search and retrieval inhibition: the role of the frontal lobes. Neuropsychologia 1993; 31: 503–24.[CrossRef][Web of Science][Medline]
Janowsky JS, Shimamura AP, Squire LR. Source memory impairment in patients with frontal lobe lesions. Neuropsychologia 1989; 27: 1043–56.[CrossRef][Web of Science][Medline]
Jetter W, Poser U, Freeman RB Jr, Markowitsch HJ. A verbal long term memory deficit in frontal lobe damaged patients. Cortex 1986; 22: 229–42.[Web of Science][Medline]
Johnson MK, Raye CL. Reality monitoring. Psychol Rev 1981; 88: 67–85.[CrossRef][Web of Science]
Johnson MK, Hashtroudi S, Lindsay DS. Source monitoring. Psychol Bull 1993; 114: 3–28.[CrossRef][Web of Science][Medline]
Kesner RP, Hopkins RO, Fineman B. Item and order dissociation in humans with prefrontal cortex damage. Neuropsychologia 1994; 32: 881–91.[CrossRef][Web of Science][Medline]
Kopelman MD. Remote and autobiographical memory, temporal context memory and frontal atrophy in Korsakoff and Alzheimer patients. Neuropsychologia 1989; 27: 437–60.[CrossRef][Web of Science][Medline]
Kopelman MD, Stanhope N, Kingsley D. Temporal and spatial context memory in patients with focal frontal, temporal lobe, and diencephalic lesions. Neuropsychologia 1997; 35: 1533–45.[CrossRef][Web of Science][Medline]
Koski LM, Paus T, Petrides M. Directed attention after unilateral frontal excisions in humans. Neuropsychologia 1998; 36: 1363–1371.[CrossRef][Web of Science][Medline]
Làdavas E, Umiltà C, Provinciali L. Hemisphere-dependent cognitive performances in epileptic patients. Epilepsia 1979; 20: 493–502.[Web of Science][Medline]
Lindsay DS, Johnson MK. The eyewitness suggestibility effect and memory for source. Mem Cogn 1989; 17: 349–358.[Web of Science][Medline]
Mangels JA. Strategic processing and memory for temporal order in patients with frontal lobe lesions. Neuropsychology 1997; 11: 207–21.[CrossRef][Web of Science][Medline]
McAndrews MP, Milner B. The frontal cortex and memory for temporal order. Neuropsychologia 1991; 29: 849–59.[CrossRef][Web of Science][Medline]
Melo B, Winocur G, Moscovitch M. False recall and false recognition: an examination of the effects of selective and combined lesions to the medial temporal lobe/diencephalon and frontal lobe structures. Cogn Neuropsychol 1999; 16: 343–59.[CrossRef][Web of Science]
Milner B. Psychological deficits produced by temporal-lobe excision. Res Publ Assoc Res Nerv Ment Dis 1958; 36: 244–57.
Milner B. Interhemispheric differences in the localization of psychological processes in man. Br Med Bull 1971; 27: 272–7.
Milner B. Disorders of learning and memory after temporal lobe lesions in man. Clin Neurosurg 1972; 19: 421–46.[Medline]
Milner B. Complementary functional specializations of the human cerebral hemispheres. In: Levi-Montalcini R, editor. Nerve cells, transmitters and behaviour. Vatican City: Pontificia Academia Scientiarum; 1980. p. 601–25.
Milner B, Corsi P, Leonard G. Frontal-lobe contribution to recency judgements. Neuropsychologia 1991; 29: 601–18.[CrossRef][Web of Science][Medline]
Mishkin M, Suzuki WA, Gadian DG, Vargha-Khadem F. Hierarchical organization of cognitive memory. Philos Trans R Soc Lond B Biol Sci 1997; 352: 1461–7.
Murray EA. Memory for objects in nonhuman primates. In: Gazzaniga MS, editor. The new cognitive neurosciences 2nd ed. Cambridge (MA): MIT Press; 2000. p. 753–63.
Olivier A. Surgery of epilepsy: methods. Acta Neurol Scand Suppl 1988; 117: 103–13.[Medline]
Petrides M. Specialized systems for the processing of mnemonic information within the primate frontal cortex. Philos Trans R Soc Lond B Biol Sci 1996; 351: 1455–62.[Web of Science][Medline]
Petrides M, Milner B. Deficits on subject-ordered tasks after frontal- and temporal-lobe lesions in man. Neuropsychologia 1982; 20: 249–62.[CrossRef][Web of Science][Medline]
Pickering AD, Mayes AR, Fairbairn AF. Amnesia and memory for modality information. Neuropsychologia 1989; 27: 1249–59.[CrossRef][Web of Science][Medline]
Ranganath C, Johnson MK, D’Esposito M. Left anterior prefrontal activation increases with demands to recall specific perceptual information. J Neurosci 2000; 20: RC108.
Schacter DL. Memory distortion: how minds, brains and societies reconstruct the past. Cambridge (MA): Harvard University Press; 1995.
Schacter DL, Harbluk JL, McLachlan DR. Retrieval without recollection: an experimental analysis of source amnesia. J Verb Learn Verb Behav 1984; 23: 593–611.[CrossRef]
Schacter DL, Norman KA, Koutstaal W. The cognitive neuroscience of constructive memory. Annu Rev Psychol 1998; 49: 289–318.[CrossRef][Web of Science][Medline]
Schwerdt PR, Dopkins S. Memory for content and source in temporal lobe patients. Neuropsychology 2001; 15: 48–57.[CrossRef][Web of Science][Medline]
Shimamura AP, Squire LR. A neuropsychological study of fact memory and source amnesia. J Exp Psychol Learn Mem Cogn 1987; 13: 464–73.[CrossRef][Web of Science][Medline]
Shimamura AP, Squire LR. The relationship between fact and source memory: Findings from amnesic patients and normal subjects. Psychobiology 1991; 19: 1–10.
Shimamura AP, Janowsky JS, Squire LE. Memory for the temporal order of events in patients with frontal lobe lesions and amnesic patients. Neuropsychologia 1990; 28: 803–13.[CrossRef][Web of Science][Medline]
Smith ML, Milner B. Effects of focal brain lesions on sensitivity to frequency of occurrence [abstract]. Soc Neurosci Abstr 1983; 9: 30.
Squire LR, Zola-Morgan S. The medial temporal lobe memory system. Science 1991; 253: 1380–6.
Squire LR, Knowlton B, Musen G. The structure and organization of memory. Annu Rev Psychol 1993; 44: 435–95.
Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain. Stuttgart: Thieme; 1988.
Tulving E. Elements of episodic memory. Oxford: Clarendon Press; 1983.
Tulving E. How many memory systems are there? Am Psychol 1985; 40: 385–98.[CrossRef]
Vargha-Khadem F, Gadian DG, Watkins KE, Connelly A, Van Paesschen W, Mishkin M. Differential effects of early hippocampal pathology on episodic and semantic memory. Science 1997; 277: 376–80.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K.-i. Yamashita, S. Hirose, A. Kunimatsu, S. Aoki, J. Chikazoe, K. Jimura, Y. Masutani, O. Abe, K. Ohtomo, Y. Miyashita, et al. Formation of Long-Term Memory Representation in Human Temporal Cortex Related to Pictorial Paired Associates J. Neurosci., August 19, 2009; 29(33): 10335 - 10340. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||