Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Mar 1.
Published in final edited form as: J Clin Exp Neuropsychol. 2019 Dec 13;42(2):171–184. doi: 10.1080/13803395.2019.1699503

The Effect of Repetition on Pronoun Resolution in Patients with Memory Impairment

Natalie V Covington 1, Jake Kurczek 2, Melissa C Duff 1, Sarah Brown-Schmidt 3
PMCID: PMC7002261  NIHMSID: NIHMS1546396  PMID: 31830861

Abstract

Referring to things in the world – that woman, her idea, she – is a central component of language. Understanding reference requires the listener to keep track of the unfolding discourse history while integrating multiple sources of information to interpret the speech stream as it unfolds in time. Pronouns are a common way to establish reference. But due to their impoverished form, to understand them listeners must relate features of the pronoun (e.g. gender, animacy) with existing representations of potential discourse referents. Successful referential processing seems to place demands on memory. In a previous study, patients with hippocampal amnesia and healthy participants listened to short stories as their eye movements were monitored. When interpreting ambiguous pronouns, healthy participants demonstrated order-of-mention effects, whereby ambiguous pronouns are interpreted as referring to the first-mentioned referent in the story. By contrast, memory-impaired patients exhibited significant disruptions in their ability to use information about which character had been mentioned first to interpret pronouns. Repetition of the most salient information is a common clinical recommendation for improving pronoun resolution and communication in individuals with memory disorders (e.g., Alzheimer’s disease) but this recommendation lacks an evidentiary basis. The present study seeks to determine whether the pronoun resolution performance of hippocampal patients can be improved, by repetition of the target referent, increasing its salience. Results indicate that patients with hippocampal damage demonstrate improved processing of pronouns following repetition of the target referent, but benefit from this repetition to a significantly smaller degree compared to healthy participants. These results provide further evidence for the role of the hippocampal-dependent memory system in language processing and point to the need for empirically tested communication interventions.

Keywords: memory, language, hippocampus, comprehension, rehabilitation

Establishing and maintaining reference is a central component of language processing, as much of what we talk about involves referring to entities in the world around us. Referential processing requires maintaining a representation of the unfolding discourse history as well as integration of multiple sources of information from the visual environment and information in speech. For example, to interpret pronouns such as she, it, and him, the listener must relate features of the pronoun (gender, animacy, etc.) with existing representations of potential discourse referents. Despite the complex calculations required to resolve referential ambiguity, healthy listeners begin this process very rapidly; within 200–400ms of pronoun onset (Arnold, Eisenband, Brown-Schmidt, & Trueswell, 2000; Kaiser, Runner, & Tanenhaus, 2009; Van Berkum, Koornneef, Otten, & Nieuwland, 2007). How this is accomplished in the brain, however, is not well understood.

Successful referential processing and resolution of referential ambiguity seems to place demands on memory. In connected speech, meaning in the moment is often derived by linking unfolding words to concepts that were previously mentioned in the discourse or that are available in the broader discourse context. This requires the listener to link an expression in the signal (e.g. “the student”, “she”) and a discourse referent that was evoked by a temporally-separated antecedent (e.g. “Kelsey”). Pronouns, in particular, are informationally impoverished forms that are interpreted by linking the referential form (e.g. “it”, “she”, “that”) to one of several candidate discourse referents that are provided by the past linguistic or non-linguistic context, a process that involves manipulating and comparing memorial representations of the unfolding language and context (Brown-Schmidt, Byron, & Tanenhaus, 2005; Heim, 1983; Webber, 2003; Kaiser & Trueswell, 2008; Clark & Sengul, 1979). Nieuwland and colleagues found a relative increase in BOLD activity in left medial temporal lobe regions for easy to resolve pronouns, compared to pronouns lacking an appropriate antecedent (Nieuwland et al., 2007). In addition, Nieuwland and Martin synthesize EEG evidence and patient results to suggest that frontotemporal and medial temporal regions work in concert to accomplish anaphor processing (Nieuwland & Martin, 2017).

A role for the medial temporal lobe in anaphor processing aligns with a proposal by Duff and Brown-Schmidt (2012), who propose that establishing and maintaining reference depends critically on and receives contributions from the hippocampal-dependent memory system. This memory system has historically been linked to the encoding and maintenance of long-term memory. More recently, hippocampal-dependent memory has been characterized less by its temporal characteristics (i.e. supporting long- versus short-term memory) as by its support of relational binding and representational flexibility (Eichenbaum & Cohen, 2001; Hannula, Tranel, & Cohen, 2006; Konkel, Warren, Duff, Tranel, & Cohen, 2008). These hallmark processing characteristics of the hippocampus allow it to bind together arbitrary co-occurrences of people, places, and things into rich representations, that become available for flexible deployment on surprisingly short time-scales (Barense, Gaffan, & Graham, 2007; Hannula & Ranganath, 2008; Hannula et al., 2006). The functionality, then, of hippocampal processing closely aligns with requirements inherent in language processing, and in the processing of reference in particular (Duff & Brown-Schmidt, 2012; 2017; MacKay et al., 2007).

Evidence for a link between referential processing and hippocampal-dependent memory comes from work with patients with hippocampal damage and amnesia. Kurczek and Duff (2011) reported that relative to healthy comparison participants, individuals with hippocampal amnesia produced fewer pronouns and anaphoric nouns in their discourse productions across genres (e.g., storytelling, picture description) and produced more incomplete anaphoric ties. These disruptions in establishing reference occurred both within and across adjacent utterances – remarkably short timescales given the historical view of hippocampal contributions to long-term memory. In addition, analysis of patient H.M.’s speech indicates use of proper names with greater-than-normal frequency relative to healthy comparison participants (Mackay et al., 2013a), which has been hypothesized to be a compensatory mechanism to offset difficulties with accurate pronoun production (MacKay et al. 2013b).

In a study of real-time language comprehension, Kurczek, Brown-Schmidt, and Duff (2013) investigated if the deficits noted in utilizing cohesive ties in language production might also be observed in an online measure of sentence comprehension. Kurczek, et al. focused their investigation on the on-line interpretation of pronouns, as interpretation of these forms involves generation of form-meaning mappings in the presence of considerable ambiguity. They hypothesized that online pronoun resolution would invoke hippocampal dependent processes. Using a visual world eye-tracking paradigm similar to one developed by Arnold and colleagues (2000), Kurczek, et al. presented participants with brief stories in which two characters were introduced, and then one was subsequently referred to with a pronoun:

Minnie is playing violin for Donald/Daisy [1]

as the sun is shining overhead. [2]

She/He is wearing a yellow bracelet. [3]

Each narrative introduced two referents [1], who were either of different gender (Minnie/Donald) or same gender (Minnie/Daisy). In all conditions, the story continued with a statement about an inanimate object that was on the screen [2]; this statement was used to draw the eyes away from the two characters. Finally, the last clause of the story [3] included the critical pronoun. The critical pronoun was unambiguous in the different-gender condition, but was ambiguous in the same-gender condition (she could refer to either Minnie or Daisy). The final two words of the sentence disambiguated the critical pronoun (e.g., across conditions, the scenes were modified such that only the target referent wore a yellow bracelet). Eye movements were monitored while participants listened to the stories. In three groups of comparison participants (healthy undergraduates, healthy older adults, and comparison patients with brain damage outside the medial temporal lobe), when the characters were of different gender, listeners quickly fixated the intended referent of the pronoun, regardless of whether that referent had been mentioned first (e.g., Minnie) or second (e.g., Donald) in the story. By contrast, when the characters were of the same gender (and therefore the pronoun was ambiguous), listeners in the comparison groups initially interpreted the pronoun as referring to the first-mentioned character, as personal pronouns typically refer to the more salient potential referent (Gundel, Hedberg, & Zacharski, 1993; Kaiser & Trueswell, 2008). This tendency to interpret an ambiguous personal pronoun as referring to the more prominent referent in the narrative (here, the first-mentioned character and subject of the sentence) has been demonstrated in multiple studies (Arnold et al., 2000; Brown-Schmidt, Byron, & Tanenhaus, 2005; Jaärvikivi, van Gompel, Hyönä, & Bertram, 2005; also see Kaiser & Trueswell, 2008; Runner, Sussman, & Tanenhaus, 2003; Foraker & McElree, 2007).

Patients with hippocampal amnesia did not significantly differ from healthy comparison participants in their use of the gender cue to map a pronoun to the matching character in the narrative. However, these patients exhibited significant disruptions in their ability to use information about which character had been mentioned first to interpret the pronoun (i.e. a reduced order of mention effect). These findings suggest that the hippocampus plays a role in maintaining and integrating information even over a very short discourse history. This disruption most likely demonstrates impairment in the ability to relate information about who was mentioned first in the story in order to distinguish between the two characters when interpreting the critical pronoun, aligning with work on hippocampal involvement in the binding of the temporal order of events (Heuer & Bachevalier, 2013; Tubridy & Davachi, 2011). Moreover, these findings suggest that persons with severe declarative memory impairment may struggle to interpret potentially ambiguous pronouns in day-to-day conversation.

The present study seeks to determine whether pronoun resolution performance of the patients with hippocampal amnesia can be “rescued” (i.e. to the level of healthy comparison participants) or improved by repetition of the target referent, increasing its salience. Repetition as a strategy for improving processing and performance has a long history in patients with profound amnesia. Repetition priming experiments demonstrate that in healthy adults, repetition of information speeds reaction times or increases accuracy, across modalities (Logan, 1990). In patients with amnesia, repetition priming of images, words, and pseudowords appears to remain intact (Cave & Squire, 1992; Smith & Oscar-Berman, 1990). Patients with amnesia demonstrated intact repetition priming for picture stimuli, lasting up to seven days, despite impaired recognition memory (Cave & Squire, 1992). Olichney, et al. (2000) examined a late positive ERP component (LPC) and the N400 (two ERP components previously shown to be sensitive to repetition in neurologically typical adults) in patients with amnesia and healthy comparison participants. Results demonstrated a beneficial effect of repetition for patients’ memory for congruous word-category + exemplar pairs, with patients’ memory ability significantly correlated with their LPC amplitude to these pairs. In contrast, in the same group of patients, these patients exhibited an intact N400 repetition reduction effect, and thus there was no relationship between N400 repetition reduction and later memory for experimental stimuli. In sum, Olichney and colleagues demonstrate differential effects of repetition on brain responses in patients with amnesia: intact N400 repetition effects (which may explain some behavioral benefits of repetition in patients with amnesia) alongside the absence of an LPC repetition effect (the amplitude of which was associated with conscious recollection in healthy comparison participants).

Further, in a study examining new semantic learning in amnesia, patient K.C. demonstrated improved semantic learning with increased repetitions of the to-be-learned stimuli (Hayman, Macdonald, & Tulving, 1993). In this study and others (see: D’Angelo et al., 2016; Freed et al., 1987; Ryan et al., 2016), patients with hippocampal amnesia may reach a particular learning “criterion” following many repetitions of a particular stimulus, compared to healthy comparison participants. In these cases, repetition as a strategy for improving processing in the moment is not the goal, but the achievement of learning to criterion suggests that repetition of to-be-learned information may be a useful strategy for improving patient performance. It should be noted, however, that reaching criterion in these studies can require repetitions several orders of magnitude larger than are feasible, or naturalistic, in real-world settings. In the current study, we test the effects of repetition in an ecologically valid way: We ask if a single repetition can improve interpretation of ambiguous pronouns.

Particularly relevant to the question of external validity is the similarity between patients with amnesia and patients with Alzheimer’s disease (AD). Hippocampal pathology and declarative memory impairment are among the earliest neuropathological signs in AD (Braak & Braak, 1991; Convit et al., 1993). Consistent with our work demonstrating links between the hippocampus and referential processing (Kurczek, Brown-Schmidt, & Duff, 2013; Kurczek & Duff, 2011), patients with AD have impairments in pronoun use and resolution (Almor, Kempler, MacDonald, Andersen, & Tyler, 1999; Dijkstra, Bourgeois, Allen, & Burgio, 2004). Such findings, along with observations of communication breakdown in patients with AD, have led to the creation of communication strategy recommendations and expert opinions to caregivers. These recommendations include verbatim repetition of the most salient information and even avoiding the use of pronouns altogether (Ripich, 1994; Ripich et al., 1998; Alzheimer’s Association, 2011; de Vries, 2013; Wilson et al., 2012; Small et al., 2003; Small & Gutman, 2002; Azuma & Bayles, 1997). While there has not been systematic investigation into the effectiveness of these recommendations, there is some empirical support for aspects of the recommended strategies. For example, Small and colleagues (1997) demonstrated that sentence comprehension improved in individuals with AD when a sentence was repeated verbatim. Furthermore, Almore and colleagues (1999) reported that language comprehension improved when individuals with AD are given full noun phrases (e.g., the young girl) instead of just a pronoun (e.g., she).

The current study examines whether similar compensation can be made by patients with hippocampal amnesia in using information about the relative salience of candidate referents when interpreting pronouns in discourse. We aim to test whether increasing the salience of the target referent by repetition might increase amnesic participants’ performance to the level of healthy comparison participants. In this study, participants listened to short narratives while looking at a related scene. Each narrative introduced two same-gender referents and then referred back to one of these referents with a pronoun. Because both potential referents were of the same gender, the pronoun was always ambiguous. The key manipulation was whether the first-mentioned referent was made more prominent through re-mention. An example narrative is presented in Table 1. If this repetition of the first-mentioned referent (increasing its salience) improves pronoun resolution in patients with hippocampal amnesia, then these results, taken in combination with results of our previous study (Kurczek et al., 2013), would suggest that the discourse representation is not entirely lost in amnesia, but rather weakened. Such a finding would also offer support for repetition as a candidate communicative intervention. In contrast, if patients with hippocampal amnesia are unable to take advantage of repetition, the finding would point to a highly degraded, or lost, representation of the discourse history and offer little support for repetition as an effective communication strategy. Regardless of outcome, results of this study will inform our understanding of hippocampal contributions to discourse representations, will provide an initial test of currently untested clinical recommendations, and will have clinical applications for populations with declarative memory impairment and hippocampal damage (e.g., Alzheimer’s disease, traumatic brain injury).

Table 1.

Example narratives by condition.

Order of mention
Target 1st mentioned Target 2nd mention
Repetition No repetition Mickey is painting a portrait of Donald and some paint is spilling on the floor. And what is he wearing? Look – he’s wearing yellow shoes Mickey is painting a portrait of Donald and some paint is spilling on the floor. And what is he wearing? Look – he’s wearing red shoes
1st mentioned character repeated Mickey is painting a portrait of Donald. Mickey is trying really hard to get the portrait just right because he wants to be a famous artist some day, and some paint is spilling on the floor. And what is he wearing? Look – he’s wearing yellow shoes Mickey is painting a portrait of Donald. Mickey is trying really hard to get the portrait just right because he wants to be a famous artist some day, and some paint is spilling on the floor. And what is he wearing? Look – he’s wearing red shoes
Open in a new tab

Target referent in bold; critical pronoun underlined; disambiguation point bold and underlined.

Methods

Participants

Four individuals (1 female, 3 male) with bilateral hippocampal damage and severe declarative memory impairment (hippocampal amnesia) and eight healthy adults participated in this study. Three participants sustained bilateral hippocampal damage from an anoxic/hypoxic event (e.g., cardiac arrest) resulting in damage thought to be restricted to the hippocampus (1846, 2363, 2563) and one participant developed herpes simplex encephalitis (HSE) (2308) resulting in extensive damage to the hippocampus, amygdala, and other MTL cortices bilaterally. Three of the patients participated in the 2013 Kurczek et al., study (1846, 2308, 2363). Patients with hippocampal amnesia were on average 56 years old (SD = 4.08) and had an average of 16 years of education (SD = 1.63). Structural MRI completed on 3 of the 4 patients confirmed bilateral hippocampal damage and volumetric analyses on 2 of the 4 patients showed significantly reduced hippocampal volumes relative to a comparison group (Table 2). All patients with bilateral hippocampal damage had a severe impairment in declarative memory (M=62.5 WMS–III GMI). All participants with hippocampal amnesia performed within normal limits on standardized neuropsychological measures of intelligence (M=95.5; WAIS–III FSIQ). Despite their severe declarative memory impairment, patients with hippocampal amnesia did not have significant disruptions on standardized neuropsychological measures of language (e.g., Boston Naming Test; Token Test), visual perception (e.g., Judgement of Line), or executive function (e.g., Wisconsin Card Sorting) (see Table 2). In addition, all four patients routinely engage in conversations, reflecting the fact that discourse remains a significant mode of communication despite their memory impairment.

Table 2.

Demographic, neuroanatomical, and neuropsychological characteristics of participants with hippocampal amnesia.

Demographic Anatomical Neuropsychological
Intelligence Memory Perception Language Executive Function
Patient Sex Hand Ed Etiology HC Vol WAIS-III FSIQ WMS-III GMI JoL BNT TT WCST PE WCST Cat
1846 F R 14 anoxia −4.23 84 57 22 43 41 6 6
2308 M L 16 HSE N/A 98 45 30 52 44 N/A N/A
2363 M R 18 anoxia −2.64 98 73 26 58 44 12 6
2563 M R 16 anoxia N/A 102 75 30 52 44 32 5
Open in a new tab

Note. M = Male; F = Female; Ed = Education; HSE= Herpes simplex Encephalitis; HC Vol = studentized residual differences in hippocampal volume relative to a matched comparison group (see Allen et al., 2006; Buchanan et al., 2005; WAIS-III = Wechsler Adult Intelligence Scale; FSIQ = Full Scale Intelligence Quotient; WMS-III GMI = Wechsler Memory Scale-III; GMI = General Memory Index; JoL = Judgement of Line Orientation Test; BNT = Boston Naming Test; TT= Token Test; WCST = Wisconsin Card Sorting Task; PE = Preservative errors; Cat = Number of categories achieved out of six.

Healthy comparison participants (n = 8, 2 female, 6 male) were matched pairwise to amnesic participants on age, sex, and level of education. Healthy comparison participants were on average 57 years old (SD = 4.82) and had an average of 17 years of education (SD = 1.51). Healthy comparison participants reported no history of neurological, psychiatric, or developmental impairments or disease.

Procedure

The participants were tested one-at-a-time in a story listening task in which they listened to a story while viewing an associated picture on a computer screen. Throughout the task their eye movements were monitored using a remote Eyelink eye-tracker (SR Research) that recorded fixation position on the screen at 1000 Hz. Each trial began with the appearance of the picture on the computer screen; the narrative began playing 3 seconds later through headphones. Following each narrative, participants selected “match” or “mismatch” to the question “Does the story match the picture?”. Each participant completed 3 practice trials followed by 128 target trials and 32 filler trials. The target and filler trials were intermingled and presented in a single random order for all participants. Patient and comparison participants completed all 163 trials in one session, with two scheduled breaks. In order to have sufficient data for statistical analysis, participants completed the task twice across two sessions (the same 163 trials were repeated across sessions), resulting in 326 total trials per participant (repetition of trials in similar paradigms was previously used with success, Kurczek, et al. 2013; Brown-Schmidt & Toscano, 2017).

Materials

Each trial consisted of a scene and a short narrative. Scenes were similar to those used in Arnold et al., 2000, Song & Fisher, 2005, and Kurczek et al., 2013. Scenes contained four familiar Disney characters (“Minnie” and “Mickey”, “Daisy” and “Donald”). There were 32 target scenes total, each of which was associated with a single picture. The 128 target trials were created by pairing the 32 target scenes to four different variants of the target narrative. Across the four variants, the critical manipulations were: 1) whether the first-mentioned character was made more salient through re-mention and 2) whether the critical pronoun referred back to the first-mentioned character or to the second-mentioned character. The narrative and the scene always matched for target trials. For filler trials 8 matched, and 24 did not match. The experimental conditions are illustrated in Table 1.

An example narrative is provided below and in Table 1. The corresponding scene showed a picture of Mickey Mouse holding a paintbrush, and Donald Duck holding a hat; a messy bucket of paint appeared in-between the two characters. In all scenes the two characters appeared on opposite sides of the screen and an inanimate object appeared between the characters. Each narrative introduced two same-gender referents [1], which establishes their initial relative salience (i.e. “Mickey” is mentioned first, and in subject position, and is therefore more salient). In the repetition conditions, the second line of the narrative [2] made the first-mentioned character more salient through re-mention; this line was not presented in the no-repetition condition.

Mickey is painting a portrait of Donald [1]

Mickey is trying really hard to get the portrait just right because he wants to be a famous artist some day [2]

and some paint is spilling on the floor [3]

And what is he wearing? [4]

Look – he’s wearing yellow shoes [5]

In all conditions, the story continued with a statement about an inanimate object that was on the screen [3]; this statement was used to draw the eyes away from the two characters. The story then continued with a rhetorical question [4], which included the critical pronoun that was the focus of our analyses. Finally, the last clause of the story [5] disambiguated the critical pronoun in the final two words of the sentence (e.g., only the target referent wore yellow shoes). Three practice trials preceded the 160 target and filler trials and were used to introduce participants to the characters. Filler trials had a similar structure to target trials, but did not contain ambiguous pronouns.

Results

Behavioral analyses

Responses to the final probe question (“Does the story match the picture?”) were used as a measure of participants’ ultimate interpretation of the critical pronoun. While all critical trials were designed such that the story matched the picture, if participants have a strong preference to interpret the pronoun as referring to the first-mentioned character (as in previous studies, Arnold et al., 2000; Kurczek et al., 2013), we would expect more “mismatch” responses when the pronoun eventually resolves to the second-mentioned character. Indeed, healthy comparison participants were highly likely to endorse the picture as matching the narrative when the target was the first-mentioned character in the no-repetition condition (M = 86.7%, SD = 28.9%), a preference that increased in the repetition condition (M = 95.2%, SD = 4.1%). When the target was the second-mentioned character, however, healthy comparison participants were considerably less likely to endorse the picture as matching in the no-repetition condition (M = 20.0%, SD = 32.8%), and even less so in the repetition condition (M = 5.7%, SD = 6.0%). Patients with hippocampal amnesia showed a different pattern. The patients endorsed the picture as matching the narrative in the first-mentioned conditions, but endorsement rates did not increase with repetition (first-mentioned, no repetition, M = 84.4%, SD = 5.9; first-mentioned, with repetition, M = 77.8%, SD = 12.4%). When the target was the second-mentioned character, patients with hippocampal amnesia were less likely to endorse the picture as matching the narrative (second-mentioned, no repetition, M = 45.3%, SD = 25.7%; second-mentioned, with repetition, M = 40.2%, SD = 31.3%).

The behavioral responses were analyzed using a logistic mixed effects model using R’s lmer function (Bates, Mächler, Bolker, & Walker, 2015). The analysis included fixed effects of Mention, Repetition, and Group, and their two-way and three-way interactions. The fixed effects were coded with mean-centered contrast codes. The dependent variable was Response (where “match” response = 1 and “mismatch” = 0). Both participants and items were treated as random intercepts. An initial model that included the maximal random effects structure justified by the design failed to converge, and a backwards-stepping procedure in which random slopes were removed from the model one-by-one in order to find the maximal model that would converge (Barr, Levy, Scheepers, & Tily, 2013). Evaluation of fixed effects was determined on the basis of this model (Table 3). Two-way interactions of Mention*Group (z = −12.02, p < 0.001), and Mention*Repetition (z = −6.77, p < 0.001) were qualified by a significant three-way interaction of Mention*Repetition*Group (z = −6.32, p < 0.001). Healthy comparison participants showed a significant effect of Mention (b = −4.59, z = −27, p < 0.001) with fewer “match” responses in the second-mentioned condition, qualified by a significant Mention*Repetition interaction (b = −2.59, z = −7.79, p < 0.001), due to a positive effect of repetition in the 1st mentioned condition (more “match” responses when the 1st mentioned character was repeatedly mentioned, and the pronoun disambiguated to the 1st mentioned character), and a negative effect of repetition in the 2nd mentioned condition (fewer “match” responses when 1st mentioned character had been repeatedly mentioned, and the pronoun disambiguated to the 2nd mentioned character).

Table 3.

Logistic mixed effects model results for analysis of behavioral data

Fixed Effects Est. SE z statistic p-value
(Intercept) 0.32 0.14 2.37 0.018
Mention [First mentioned = −0.5, second mentioned = 0.5] −3.67 0.12 −29.62 < 0.001
Repetition [No repetition = −0.5, repetition = 0.5] −0.21 0.12 −1.72 0.086
Group [Patients = −0.67, Comparisons = 0.33] −9.46 0.28 −1.62 0.105
Mention*Repetition −1.64 0.24 −6.77 < 0.001
Mention*Group −2.73 0.23 −12.02 < 0.001
Repetition*Group 0.20 0.22 0.88 0.378
Mention*Repetition*Group −2.82 0.45 −6.32 < 0.001
Random Effects Variance SD  
Participant (Intercept) 0.18 0.42
Observations: 3031; Participants: 12
Open in a new tab

Note. The dependent measure is a binary measure of whether the participant responded that the story did (1) or did not (0) match the picture. Bolded values indicate statistically significant results. Contrast codes for fixed effects reported in brackets.

By contrast, in the amnesic group the Mention*Repetition interaction was not significant (b = 0.23, z = 0.77, p = 0.44). The amnesic group demonstrated a significant, but attenuated, main effect of Mention (b = −1.87, z = −12.33, p < 0.001), with fewer “match” responses when the target was the second-mentioned character, and a significant main effect of Repetition (b = −0.34, z = −2.27, p = 0.02) with overall fewer “match” responses when the first-mentioned referent was repeated. Full model results are presented in Table 3.

Eye-gaze analyses

Data prior to the first break (i.e. one third of the trials) from one healthy comparison (1846C-1) were lost due to a technical error and were not included in data analysis. A plot of eye fixations over time is presented in Figure 1. The primary measure of interest was the eye fixations that participants made to the target and competitor referents as they interpreted the critical pronoun. In a sentence like “Mickey is painting a portrait of Donald…”, if the critical pronoun resolved to the 1st-mentioned character, fixations to the picture of Mickey were categorized as target fixations, and fixations to the picture of Donald were categorized as competitor fixations (and vice-versa for pronouns that resolved to the 2nd-mentioned character). For each trial, the proportion of fixations to the target and competitor referents was calculated for the time window 200 ms to 1000 ms following onset of the critical pronoun. This time window was offset by 200 ms in order to account for the time required to program and launch an eye movement (Hallett, 1986). Figure 1 shows that during interpretation of the pronoun the preference to fixate the target over the competitor generally rises in the 1st mentioned condition, and falls in the 2nd mentioned condition, reflecting the preference to interpret the pronoun as referring to the more salient of the two candidate referents. This pattern is clearly attenuated in the patient group. In addition, we note that prior to the onset of the pronoun (left side of figure) the healthy comparison participants show a larger target advantage score in the 1st mentioned condition compared to the 2nd mentioned condition. This pattern, which is absent in the patient group, likely reflects anticipation of the story as continuing about the first-mentioned referent, consistent with maintenance of a representation of the discourse structure over time.

Figure 1.

Figure 1.

Open in a new tab

Target advantage (proportion of fixations to the target referent minus proportion of fixations to the competitor referent) across the four experimental conditions for healthy comparison participants (left panel) and participants with amnesia (right panel). Time is on the x-axis in milliseconds; 0 ms corresponds to the onset of the critical pronoun (… And what is she wearing?…). Vertical dotted lines indicate the analysis window, offset 200 ms from pronoun onset.

For statistical analysis of eye movement data, the dependent measure was the log of the ratio of target to competitor fixations (LogRatioTC; following Kurczek, et al., 2013), calculated separately for each trial. To handle trials in which the proportion of fixations to the competitor was zero, and the ratio is undefined, 0.5 was added to both numerator and denominator (see a similar approach in Barr, 2008; Heller, Grodner, & Tanenhaus, 2008). A plot of LogRatioTC by condition is presented in Figure 2. Positive values of the dependent variable indicate a target preference, and negative values indicate a competitor preference; this measure is preferred over simply analyzing target fixations because it captures the degree of competition between the two animate referents.

Figure 2.

Figure 2.

Open in a new tab

Plot of the primary dependent measure (LogRatioTC) by condition. Boxplots represent group medians and interquartile range. Points indicate individual participant averages.

The data were analyzed using a linear mixed effect model with LogRatioTC as the dependent variable1, and fixed effects of Group, Mention, Repetition and their interactions. The fixed effects were coded with mean-centered contrast codes. As before, participants and items were included as random intercepts, and a backwards-stepping procedure was used to identify the model with the maximal number of random slopes that converged. The full model results including both fixed and random effects are presented in Table 4. This analysis revealed significant two-way interactions of Mention*Repetition and Mention*Group that were qualified by a significant three-way Group*Mention*Repetition interaction (t = −3.40, p < 0.001). These findings demonstrate that patients with hippocampal amnesia and healthy comparison participants were differentially sensitive to order of mention cues and the repetition manipulation. In particular, healthy comparison participants exhibited a significant main effect of Mention (b = −8.71, t = −7.47, p < 0.001), with fewer looks to the target in the second-mentioned condition. This main effect was qualified by a significant Mention*Repetition interaction (b = −3.57, t = −9.22, p < 0.001). When the pronoun would eventually resolve to the 1st mentioned character (e.g. Mickey…. Donald… he-> Mickey), comparison participants were more likely to look at the target when that character had been repeatedly mentioned (M = 5.35), compared to the no-repetition condition (M = 3.50). The opposite effect was observed when the pronoun would eventually resolve to the 2nd mentioned character – in this case, comparison participants were more likely to look at the competitor (the 1st-mentioned character) (M = −3.40), and repeated mention of the 1st-mentioned character caused participants to be even more likely to look at the competitor (M = −5.11) (the 1st-mentioned character), more than the target (the 2nd mentioned character).

Table 4.

Mixed effects model results for analysis of eye-gaze data: Critical Pronoun

Fixed Effects Est. SE t statistic p-value
(Intercept) 0.01 0.11 2.37 0.905
Mention [First mentioned = −0.5, second mentioned = 0.5] −7.13 0.98 −7.26 < 0.001
Repetition [No repetition = −0.5, repetition = 0.5] −0.02 0.17 −0.13 0.899
Group [Patients = −0.67, Comparisons = 0.33] 0.28 0.29 0.97 0.343
Mention*Repetition −2.82 0.32 −8.86 < 0.001
Mention*Group −4.88 2.03 −2.40 0.036
Repetition*Group 0.29 0.36 0.80 0.427
Mention*Repetition*Group −2.31 0.68 −3.40 < 0.001
Random Effects Variance SD
ID (Intercept) 0.07 0.27
Group 1.11 1.06
Mention 2.57 1.60
Group:Mention 4.93 2.22
Participant (Intercept) 0.05 0.22
Mention 10.30 3.21
Repetition 0.03 0.18
Residual 18.72 4.33
Observations: 2962; Participants: 12; ID: 32
Open in a new tab

Note. The dependent measure is the log odds of target to competitor fixations during the 200 to 1000 ms following critical pronoun onset. Bolded values indicate statistically significant results. Contrast codes for fixed effects reported in brackets.

Patients with hippocampal amnesia demonstrated a similar, but markedly attenuated, pattern of effects: A significant main effect of Mention (b = −3.83, t = 2.25, p = 0.04) was qualified by a significant, but smaller Mention*Repetition interaction (b = −1.26, t = −2.25, p = 0.02). When the target was the 1st-mentioned referent, patients demonstrated a target preference (M = 1.53) that increased somewhat when the 1st mentioned referent was repeated (M = 1.97), but remained numerically smaller than comparison participants. When the pronoun would eventually resolve to the 2nd mentioned character, patients were more likely to look at the competitor (the 1st mentioned character; M = −1.78), an effect that was amplified by repetition of the 1st-mentioned character (M = −2.59) but numerically smaller than healthy comparisons. Full model results are presented in Table 4.

To ensure that group differences in repetition and order-of-mention effects are not the result of gross differences between groups in the ability to move the eye in response to a clear external stimulus, an equivalent analysis was run at the disambiguation point – the point at which the listener explicitly learns which character is the target referent (e.g. … she’s wearing yellow shoes.). Full model results are presented in Table 7. Critically, this analysis did not reveal any significant interactions with group (ps>.11). Instead, at the disambiguation point there was a significant main effect of Mention (t = −5.79, p < 0.001), with fewer looks to the target in the second-mentioned condition. This main effect was qualified by a significant interaction of Mention*Repetition (t = −2.80, p = 0.005), such that repetition of the first-mentioned character had a positive effect on looks to the target referent in the first-mentioned condition and a negative effect on looks to the target referent in the second-mentioned condition.

Table 7.

Full model output for analysis of eye-gaze data: Disambiguation Point

Fixed Effects Est. SE t statistic p-value
(Intercept) 0.66 0.24 2.75 0.016
Mention [First mentioned = −0.5, second mentioned = 0.5] −6.06 1.05 −5.79 < 0.001
Repetition [No repetition = −0.5, repetition = 0.5] −0.05 0.13 −0.38 0.706
Group [Patients = −0.67, Comparisons = 0.33] −0.17 0.48 −0.36 0.727
Mention*Repetition −0.75 0.27 −2.80 0.005
Mention*Group −3.67 2.15 −1.70 0.119
Repetition*Group 0.14 0.29 0.48 0.630
Mention*Repetition*Group −0.48 0.57 −0.84 0.402
Random Effects Variance SD
ID (Intercept) 0.26 0.51
Group 0.25 0.50
Mention 2.63 1.62
Group:Mention 2.30 1.52
Participant (Intercept) 0.54 0.73
Mention 11.96 3.46
Residual 13.27 3.64
Observations: 2951; Participants: 12; ID: 32
Open in a new tab

Note. The dependent measure is the log odds of target to competitor fixations during the 200 to 1000 ms following the disambiguation point. Bolded values indicate statistically significant results. Contrast codes for fixed effects reported in brackets.

Study data are publicly available on the Open Science Framework at https://osf.io/3u5fm/; study materials are available by request.

Discussion

The current study builds upon previous findings demonstrating that patients with declarative memory impairment struggle to resolve ambiguous pronouns during online language processing. We sought to determine whether the performance of the patients with hippocampal amnesia could be rescued to the level of healthy comparison participants, or improved, by repeated mentioning of the first-mentioned character in the story prior to the ambiguous pronoun, a common but empirically untested clinical recommendation for communicating with memory-impaired individuals. Consistent with previous findings (Kurczek et al., 2013; MacKay et al., 2007) and our proposal on the role of the hippocampus in language processing (Duff & Brown-Schmidt, 2012, 2017), the present results demonstrate that patients with profound memory deficits experience difficulty integrating and maintaining referential information over short discourse histories. Increasing the salience of the first-mentioned referent via repetition supported interpretation of pronouns as referring to that character, but this effect was attenuated compared to healthy comparison participants and did not raise patients’ performance to the level of healthy individuals.

Healthy comparison participants demonstrated a strong order-of-mention effect that was amplified by the repetition manipulation: when the first-mentioned character had been repeatedly mentioned, healthy participants were more likely to look at the target when the intended referent was the 1st-mentioned character, but significantly less likely to look at the target when the intended referent was the 2nd-mentioned character. This indicates that, for healthy comparison participants, repetition successfully increased the salience of the 1st mentioned character, making it a stronger candidate referent of the pronoun (see Grosz, et al., 1995; Gordon, et al., 1993; Song & Fisher, 2005). By contrast, although patients with hippocampal amnesia demonstrated a significant order-of-mention effect, it was attenuated relative to the healthy comparison participants. Further, while effects of repetition maintained the same directionality in patients with hippocampal amnesia, these effects were much smaller. It should be noted that while the present study found a significant (albeit smaller) order-of-mention effect in the patients with hippocampal amnesia, the original study by Kurczek et al. (2013) did not find a significant effect of order-of-mention in patients with hippocampal amnesia. Differences in the patient group included in the current study may be in play. The current study included patient 2563 instead of patient 1951 from the Kurczek study. While both patients have profound declarative memory impairments relative to healthy adults, 1951 scores nearly two standard deviations lower than 2563 on the WMS (57 versus 75). Another potentially relevant difference may be that the present study tested pronoun interpretation in narratives where the candidate referents were always of the same gender. In contrast, Kurczek et al. (2013) included narratives with different-gender characters, which may have lessened the need to attend to order-of-mention information. Taken together, results from these two studies suggest that patients with hippocampal amnesia maintain some representation of the discourse structure, but are significantly impaired relative to individuals without hippocampal damage and profound memory impairment.

Comprehension of pronouns during discourse processing requires flexible integration of multiple sources of information on a rapid timescale. The current study sought to bolster the amount of information available to patients with hippocampal amnesia by increasing the salience of the target referent via repetition – by making it the clear topic of the story. Stories with repeated mention of the 1st mentioned character were also longer, allowing more time to form a strong representation of the unfolding discourse structure. While healthy comparison participants were able to make use of this additional information, memory-impaired patients did not benefit to the same extent. Thus, it appears that the same challenges in integrating and maintaining referential information over the course of a single mention extends to the repetition of that same content.

A common and oft-cited recommendation for improving communication with memory-impaired individuals (e.g. individuals with AD, TBI, etc.) is repetition of important information, repeated either verbatim or in paraphrase (Ripich, 1994; Ripich et al., 1998; Alzheimer’s Association, 2011; de Vries, 2013; Wilson et al., 2012; Small et al., 2003; Small & Gutman, 2002; Azuma & Bayles, 1997). While these recommendations are prevalent in clinical practice and in guidelines set forth by patient advocacy organizations, they have largely remained untested by empirical studies (but see: Almore et al., 1999). Here, we find empirical evidence that, at least for patients with profound declarative memory impairment, repetition of salient information provides only a modest improvement in comprehension and does not raise performance to the level of healthy comparison participants. An additional finding with translational relevance is that patients, like healthy participants, struggled to interpret pronouns referencing the 2nd mentioned of the two characters. This finding suggests that use of pronouns to refer to less salient entities may present comprehension challenges for persons with severe memory impairment.

Our findings point to the critical need to empirically evaluate current communication strategy recommendations, rather than relying solely on expert opinion or “common sense” recommendations. Without a strong research foundation, dissemination of such recommendations may result in wasted efforts towards strategies that are not effective and wasted financial and time resources devoted to training family members and staff. Such recommendations may also lead to false hope or unrealistic expectations by caregivers. While the current findings demonstrate some benefit of repetition for pronoun comprehension, repetition did not produce normal or intact performance, suggesting that individuals with memory impairment would continue to struggle in resolving ambiguous pronouns in conversation, even with repetition. Future research should also address the impact of recommended communication strategies in multiparty interactions (e.g., interactions between individuals with memory impairment and their unimpaired caregivers). For example, repeating a name (when you would typically use a pronoun) is known to disrupt processing and understanding in healthy individuals (Gordon et al., 1993). (Note this is a different repetition manipulation than we studied here; we repeated the name where a pronoun would have been ambiguous and nominalized it prior to the critical pronoun). Thus, research on the effectiveness of a given communication strategy should examine the distributed impact across communication partners (e.g., Duff et al., 2011), as the effects of different clinical recommendations may have differential impacts on impaired versus healthy conversation partners.

In the current study, we tested the strategy of repeating information in discourse to increase its salience and support subsequent pronoun comprehension. Other communication strategies to improve pronoun resolution include verbatim repetition and avoidance of pronouns altogether. Research directed at determining the effectiveness of such strategies in improving comprehension and communicative outcomes should be carried out in populations for whom these recommendations are directed (e.g. AD). The effect of repetition observed here was in individuals with relatively “pure” memory deficits (e.g. in the absence of deficits in other cognitive domains such as attention, executive functioning, etc.) and with focal lesions (in contrast with more diffuse pathology implicated in AD and TBI). The fact that we observed only small effects of repetition leads to the prediction that repetition may be unlikely to benefit comprehension in more complex disorders (e.g. AD, TBI). It remains an open question as to whether the current repetition manipulation would improve comprehension in individuals with less severe memory impairment. There is a clear need for more replication studies and empirical tests of current communication strategies and interventions in individuals with cognitive-communication disorders and for the development of new theoretically motivated treatments to improve the communicative lives of individuals with memory disorders.

The current study is limited by its small sample size, a result of the rarity of patients with focal medial temporal lobe damage and profound amnesia. To address issues of statistical power, participants completed a large number of critical trials, determined based upon previous studies in our lab and others (Kurczek et al., 2013; Arnold et al., 2000; Brown-Schmidt & Toscano, 2017). These results provide an initial test of repetition as a candidate compensatory strategy for patients with memory impairment, in patients with relatively “clean” neuropsychological profiles. These results should be extended to larger groups of patients with memory impairment caused by more frequently occurring etiologies (i.e. AD).

In conclusion, the present research tracked the eye movements of patients with hippocampal amnesia and healthy comparison participants as they listened to short stories involving an ambiguous pronoun. While healthy comparisons used the structure of these narratives to interpret ambiguous pronouns, patients with hippocampal amnesia showed only limited use of the same sources of information, and never achieved similar levels of performance as healthy participants. These findings suggest that memory-impaired individuals do not show large benefits from repetition as a strategy to increase the salience of a potentially ambiguous pronoun and demonstrate that clinical recommendations proposing repetition as a strategy to improve language comprehension, while not harmful, may provide only limited benefit.

Figure 3.

Figure 3.

Open in a new tab

Plot of behavioral “Match” responses by condition. Boxplots represent group medians and interquartile range. Points indicate individual participant averages.

Figure 4.

Figure 4.

Open in a new tab

Plot of the secondary dependent measure (proportion of trials with at least one look to target) by condition. Boxplots represent group medians and interquartile range. Points indicate individual participant averages.

Figure 5.

Figure 5.

Open in a new tab

Plot of LogRatioTC at the disambiguation point, by condition. Boxplots represent group medians and interquartile range. Points indicate individual participant averages.

Table 5.

Logistic mixed effects model results for analysis of eye gaze: Critical Pronoun

Fixed Effects Est. SE z statistic p-value
(Intercept) 0.64 0.13 4.80 < 0.001
Mention [First mentioned = −0.5, second mentioned = 0.5] −2.68 0.44 −6.16 < 0.001
Repetition [No repetition = −0.5, repetition = 0.5] 0.09 0.10 0.87 0.383
Group [Patients = −0.67, Comparisons = 0.33] 0.37 0.24 1.55 0.122
Mention*Repetition −1.11 0.20 −5.56 < 0.001
Mention*Group −2.31 0.88 −2.64 0.008
Repetition*Group 0.23 0.19 1.18 0.238
Mention*Repetition*Group −1.33 0.39 −3.43 < 0.001
Random Effects Variance SD
ID (Intercept) 0.15 0.38
Mention 0.51 0.71
Participant (Intercept) 0.12 0.35
Mention 1.93 1.39
Observations: 2963; Participants: 12; ID: 32
Open in a new tab

Note. The dependent measure is a binary measure of whether the participant looked at the target (1) or not (0) during the 200 to 1000 ms following critical pronoun onset. Bolded values indicate statistically significant results. Contrast codes for fixed effects reported in brackets.

Table 6.

Model-predicted probability of target fixation during interpretation of the critical pronoun (200 to 1000 ms following pronoun onset)

Healthy Comparisons Amnesic Patients
No Repetition First-Mentioned 0.88 0.72
Second-Mentioned 0.34 0.48
Repetition First-Mentioned 0.95 0.73
Second-Mentioned 0.22 0.44
Open in a new tab

Acknowledgements:

The authors thank Sharice Clough for her help in data collection. Funding Details: This work was supported by NIDCD grant R01 DC011755 to MCD and SBS.

Footnotes

Disclosure of Interest: Research funding from NIDCD grant R01 DC011755. The authors have no significant financial or other conflicts of interest to report.

1

LogRatioTC was chosen as the dependent variable to maintain consistency with the previous study (Kurczek et al., 2013). However, LogRatioTC has been noted to be non-normally distributed. Therefore, following Ryskin, et al., (2015), a supplemental analysis used a binary dependent variable, with 1 = target fixation during the trial; 0 = no target fixation. Results of this analysis were equivalent to the LogRatioTC analysis and are reported in Table 5. In addition, model-predicted probabilities of target fixations in the critical region of analysis are presented in Table 6.

References

  1. Allen JS, Tranel D, Bruss J, & Damasio H (2006). Correlations between Regional Brain Volumes and Memory Performance in Anoxia. Journal of Clinical and Experimental Neuropsychology, 28(4), 457–476. [DOI] [PubMed] [Google Scholar]
  2. Almor A, Kempler D, MacDonald MC, Andersen ES, & Tyler LK (1999). Why do Alzheimer patients have difficulty with pronouns? Working memory, semantics, and reference in comprehension and production in Alzheimer’s disease. Brain and Language, 67(3), 202–227. [DOI] [PubMed] [Google Scholar]
  3. Arnold JE, Eisenband JG, Brown-Schmidt S, & Trueswell JC (2000). The rapid use of gender information: Evidence of the time course of pronoun resolution from eyetracking. Cognition, 76(1), 13–26. [DOI] [PubMed] [Google Scholar]
  4. Azuma T, & Bayles K (1997). Memory Impairments Underlying Language Difficulties in Dementia. Topics in Language Disorders, 18(1), 58–71. [Google Scholar]
  5. Baayen RH, Davidson DJ, & Bates DM (2008). Mixed-effects modeling with crossed random effects for subjects and items. Journal of memory and language, 59(4), 390–412. [Google Scholar]
  6. Barense MD, Gaffan D, & Graham KS (2007). The human medial temporal lobe processes online representations of complex objects. Neuropsychologia, 45(13), 2963–2974. [DOI] [PubMed] [Google Scholar]
  7. Barr DJ (2008). Analyzing ‘visual world’ eyetracking data using multilevel logistic regression. Journal of memory and language, 59(4), 457–474. [Google Scholar]
  8. Barr DJ, Levy R, Scheepers C, & Tily HJ (2013). Random effects structure for confirmatory hypothesis testing: Keep it maximal. Journal of memory and language, 68(3), 255–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bates D, Mächler M, Bolker B, & Walker S (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. [Google Scholar]
  10. Braak H, & Braak E (1991). Neuropathological stageing of Alzheimer-related changes. Acta Neuropathologica, 82(4), 239–59. [DOI] [PubMed] [Google Scholar]
  11. Brown-Schmidt S, Byron DK, & Tanenhaus MK (2005). Beyond salience: Interpretation of personal and demonstrative pronouns. Journal of Memory and Language, 53(2), 292–313. [Google Scholar]
  12. Brown-Schmidt S, & Toscano JC (2017). Gradient acoustic information induces long-lasting referential uncertainty in short discourses. Language, Cognition and Neuroscience, 32(10), 1211–1228. [Google Scholar]
  13. Buchanan TW, Tranel D, & Adolphs R (2005). Emotional Autobiographical Memories in Amnesic Patients with Medial Temporal Lobe Damage. The Journal of Neuroscience, 25(12), 3151–3160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Clark HH, & Sengul CJ (1979). In search of referents for nouns and pronouns. Memory & Cognition, 7(1), 35–41. [Google Scholar]
  15. Convit A, de Leon MJ, Golomb J, George AE, Tarshish CY, Bobinski M, … Wisniewski H (1993). Hippocampal atrophy in early Alzheimer’s Disease: Anatomic specificity and validation. Psychiatric Quarterly, 64(4), 371–387. [DOI] [PubMed] [Google Scholar]
  16. D’Angelo MC, Rosenbaum RS, & Ryan JD (2016). Impaired inference in a case of developmental amnesia. Hippocampus, 26(10), 1291–1302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. De Vries K (2013). Communicating with older people with dementia. Nursing older people, 25(4), 30–37. [DOI] [PubMed] [Google Scholar]
  18. Dijkstra K, Bourgeois MS, Allen RS, & Burgio LD (2004). Conversational coherence: Discourse analysis of older adults with and without dementia. Journal of Neurolinguistics, 17(4), 263–283. [Google Scholar]
  19. Duff MC, & Brown-Schmidt S (2012). The hippocampus and the flexible use and processing of language. Frontiers in Human Neuroscience, 6(April), 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Eichenbaum H, & Cohen N (2001). From conditioning to conscious recollection: Memory systems of the brain. New York: Oxford University Press. [Google Scholar]
  21. Foraker S, & McElree B (2007). The role of prominence in pronoun resolution: Active versus passive representations. Journal of Memory and Language, 56(3), 357–383. [Google Scholar]
  22. Freed DM, Corkin S, & Cohen NJ (1987). Forgetting in HM: A second look. Neuropsychologia, 25(3), 461–471. [DOI] [PubMed] [Google Scholar]
  23. Gerrig RJ, & McKoon G (1998). The readiness is all: The functionality of memory-based text processing. Discourse Processes, 26, 67–86. [Google Scholar]
  24. Gundel JK, Hedberg N, & Zacharski R (1993). Cognitive Status and the Form of Referring Expression in Discourse. Language, 69(2), 274–307. [Google Scholar]
  25. Hallett P (1986). Eye movements and human visual perception. Handbook of perception and human performance., 1, 10–1. [Google Scholar]
  26. Hannula D, & Ranganath C (2008). Medial temporal lobe activity predicts successful relational memory binding. Journal of Neuroscience, 28, 116–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hannula D, Tranel D, & Cohen NJ (2006). The Long and the Short of It: Relational Memory Impairments in Amnesia, Even at Short Lags. Journal of Neuroscience, 26(32), 8352–8359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Heim I (1983). File change semantics and the familiarity theory of definiteness In Bäuerle R, Schwarze C, & Stechow A. v. (Eds.), Meaning, use, and the interpretation of language. Berlin: de Gruyter. [Google Scholar]
  29. Heller D, Grodner D, & Tanenhaus MK (2008). The role of perspective in identifying domains of reference. Cognition, 108(3), 831–836. [DOI] [PubMed] [Google Scholar]
  30. Heuer E, Bachevalier J. (2013). Working memory for temporal order is impaired after selective neonatal hippocampal lesions in adult rhesus macaques. Behavioral Brain Research, 239, 55–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jaeger TF (2008). Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed models. Journal of memory and language, 59(4), 434–446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Järvikivi J, van Gompel RP, Hyönä J, & Bertram R (2005). Ambiguous Pronoun Resolution: Contrasting the First-Mention and Subject Preference Accounts. Psychological Science, 16, 260–264. [DOI] [PubMed] [Google Scholar]
  33. Kaiser E, Runner JT, & Tanenhaus MK (2009). Structural and semantic constraints on the resolution of pronouns and reflexives. Cognition, 112(1), 55–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kaiser E, & Trueswell JC (2008). Interpreting pronouns and demonstratives in Finnish: Evidence for a form-specific approach to reference resolution. Language and Cognitive Processes, 23(5), 709–748. [Google Scholar]
  35. Konkel A, Warren DE, Duff MC, Tranel DN, & Cohen NJ (2008). Hippocampal amnesia impairs all manner of relational memory. Frontiers in Human Neuroscience, 2(October), 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kurczek J, Brown-Schmidt S, & Duff M (2013). Hippocampal contributions to language: evidence of referential processing deficits in amnesia. Journal of Experimental Psychology. General, 142(4), 1346–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Kurczek J, & Duff MC (2011). Cohesion, coherence, and declarative memory: Discourse patterns in individuals with hippocampal amnesia. Aphasiology, 25(6–7), 700–712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. MacKay, James, Taylor, & Marian (2007). Amnesic H.M. exhibits parallel deficits and sparing in language and memory: Systems versus binding theory accounts. Language and Cognitive Processes, 22(3):377–452. [Google Scholar]
  39. MacKay, Johnson, Fazel & James (2013a). Compensating for Language Deficits in Amnesia I: H.M.’s Spared Retrieval Categories. Brain Sciences, 3(1):262–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. MacKay, Johnson, & Hadley (2013b). Compensating for Language Deficits in Amnesia II: H.M.’s Spared versus Impaired Encoding Categories. Brain Sciences, 3(2):415–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Nieuwland MS, & Martin AE (2017). Neural oscillations and a nascent corticohippocampal theory of reference. Journal of Cognitive Neuroscience, 29(5), 896–910. [DOI] [PubMed] [Google Scholar]
  42. Olichney JM, Van Petten C, Paller KA, Salmon DP, Iragui VJ, & Kutas M (2000). Word repetition in amnesia: Electrophysiological measures of impaired and spared memory. Brain, 123(9), 1948–1963. [DOI] [PubMed] [Google Scholar]
  43. Ripich DN (1993). Functional communication with AD patients: a caregiver training program. Alzheimer Disease and Associated Disorders, 8, 95–109. [PubMed] [Google Scholar]
  44. Ripich DN, Ziol E, & Lee MM (1998). Longitudinal effects of communication training on caregivers of persons with Alzheimer’s disease. Clinical Gerontologist, 19(2), 37–55. [Google Scholar]
  45. Ryan JD, D’Angelo MC, Kamino D, Ostreicher M, Moses SN, & Rosenbaum RS (2016). Relational learning and transitive expression in aging and amnesia. Hippocampus, 26(2), 170–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Small JA, & Gutman G (2002). Recommended and reported use of communication strategies in Alzheimer caregiving. Alzheimer Disease & Associated Disorders, 16(4), 270–278. [DOI] [PubMed] [Google Scholar]
  47. Small JA, Gutman G, Makela S, & Hillhouse B (2003). Effectiveness of communication strategies used by caregivers of persons with Alzheimer’s disease during activities of daily living. Journal of Speech, Language, and Hearing Research, 46(2), 353–367. [DOI] [PubMed] [Google Scholar]
  48. Smith ME, & Oscar-Berman M (1990). Repetition priming of words and pseudowords in divided attention and in amnesia. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16(6), 1033. [DOI] [PubMed] [Google Scholar]
  49. Song HJ, & Fisher C (2005). Who’s “she”? Discourse prominence influences preschoolers’ comprehension of pronouns. Journal of Memory and Language, 52(1), 29–57. [Google Scholar]
  50. Tubridy S, Davachi L (2011). Medial temporal lobe contributions to episodic sequence encoding. Cerebral Cortex, 21(2), 272–280. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Van Berkum JJA, Koornneef AW, Otten M, & Nieuwland MS (2007). Establishing reference in language comprehension: An electrophysiological perspective. Brain Research, 1146(1), 158–171. [DOI] [PubMed] [Google Scholar]
  52. Webber B, Stone M, Joshi A, & Knott A (2003). Anaphora and discourse structure. Computational linguistics, 29(4), 545–587. [Google Scholar]
  53. Wilson R, Rochon E, Mihailidis A, & Leonard C (2012). Examining success of communication strategies used by formal caregivers assisting individuals with Alzheimer’s disease during an activity of daily living. Journal of Speech, Language, and Hearing Research, 55(2), 328–341 [DOI] [PubMed] [Google Scholar]

RESOURCES