Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Oct 21.
Published in final edited form as: Brain Lang. 2019 Nov 15;201:104711. doi: 10.1016/j.bandl.2019.104711

The hippocampus and semantic memory over time

Nathaniel B Klooster a,b, Daniel Tranel c,*, Melissa C Duff d
PMCID: PMC7577377  NIHMSID: NIHMS1565198  PMID: 31739112

Abstract

We previously reported impoverished semantic memory in patients with hippocampal amnesia (Klooster & Duff, 2015). Here, we test whether this disruption results from the patients not updating semantic representations since the onset of their amnesia. We extend previous work by comparing performance of hippocampal patients and their current age (CA) comparisons (M = 58.5 years) to a new comparison group matched to the patients’ age of onset (AoO) of hippocampal damage (M = 36.8). Participants completed feature and senses-listing tasks and the Word Associates Test. Both comparison groups performed significantly better than the patients with amnesia. A key new finding was that the older CA group performed significantly better than the younger AoO group. Semantic memory may become richer over time as additional information is added to existing representations. We conclude that a failure to update semantic memory may explain (at least some of) the previously observed deficits in amnesia and that the hippocampus may support semantic memory across the lifespan. Longitudinal data from patients with hippocampal pathology would provide a critical test of our conclusion.

Keywords: Semantic memory, Semantic richness, Hippocampus, Consolidation, Remote memory

1. Introduction

The acquisition of semantic memory (e.g., knowledge of vocabulary and general information about the world and oneself (Tulving, 1972)) has long been linked to the hippocampus, and medial temporal lobes, through their role in relational binding (i.e., the linking of the arbitrary relations between a concept and its label (Eichenbaum & Cohen, 2001; Manns, Hopkins, & Squire, 2003; Postle & Corkin, 1998; see Duff, Covington, Hilverman, & Cohen, in press, for review)). Over time, semantic knowledge is thought to become independent of the hippocampus through neocortical consolidation (McClelland, McNaughton, & O’reilly, 1995; O’Reilly & Rudy, 2000). Some of the most compelling evidence for this view comes from patients with hippocampal amnesia who have severe deficits in acquiring new vocabulary (Gabrieli, Cohen, & Corkin, 1988; Manns et al., 2003), but perform within normal limits on standardized tests of remote semantic knowledge, information acquired long before the onset of their amnesia (Kensinger, Ullman, & Corkin, 2001; Manns et al., 2003; Verfaellie, Koseff, & Alexander, 2000). Such findings led to the view that the contribution of the hippocampus to semantic memory is limited to acquisition and that remote semantic memory is fully intact in amnesia.

In recent work, however, we reported disruptions in the depth and richness of remote semantic knowledge in a group of patients with hippocampal damage and amnesia (Klooster & Duff, 2015). The previous work on remote semantic memory in amnesia focused on surface level knowledge – requiring participants to name a picture of an object, to match a label with a short definition, or pick a definition out of a list of foils (i.e. pen: a writing utensil that uses ink). A critical difference in our work is that we assessed semantic richness and depth of knowledge including the number of features associated with a given concept (i.e. pen: has ink, has a ballpoint, used for signing checks, can be erasable, can be permanent, has a cap, held in pockets), the number of senses a word can take (pen: a writing instrument; to write a letter; an animal enclosure; to keep in an enclosure; a female swan), and the associates and collocates of a word (e.g., words that frequently co-occur: pencil; paper; fountain; tip; letter, knife). When remote semantic knowledge was assessed at this depth, a common approach in the psycholinguistics and language-learning literatures, the amnesic patients had significantly fewer features, senses, and associates linked to their concepts than both healthy and brain-damaged comparison participants matched to the patients’ age at testing. Here, we investigate a potential mechanism for the previously observed deficits.

We interpreted the deficits in remote semantic knowledge in patients with hippocampal amnesia as a direct consequence of their hippocampal damage rather than a consequence of brain damage more generally (i.e., a group of patients with ventromedial prefrontal cortex (vmPFC) damage performed like healthy participants) (Klooster & Duff, 2015). More specifically, we proposed that, consistent with views of word learning as a protracted process unfolding over days, weeks, months, and even years (Carey, 2010; McMurray, Horst, & Samuelson, 2012), the hippocampus, in its capacity for relational binding (Cohen & Eichenbaum, 1993; Eichenbaum & Cohen, 2001), reconsolidation (Lee, 2008; McKenzie & Eichenbaum, 2011), and flexible integration of information (Zeithamova & Preston, 2010), is well situated to meet the demands of maintaining and updating semantic memory as more information is associated with a word or concept over time and with experience (e.g., McGregor, Newman, Reilly, & Capone, 2002; Laszlo & Federmeier, 2011). From this perspective, the hippocampus may support remote semantic knowledge by strengthening and enriching existing semantic representations through the addition and integration of new information (e.g., features or senses of a word) as words are retrieved and used in novel contexts in the negotiation of meaning in the moment.

Vocabulary size and the breadth of semantic knowledge are among some of the only human faculties thought to remain stable, and even improve with age (Light, 1991; Metcalfe, Casal-Roscum, Radin, & Friedman, 2015; Ramscar, Hendrix, Shaoul, Milin, & Baayen, 2014; Salthouse, 1993; Staudinger, Cornelius, & Baltes, 1989; Umanath & Marsh, 2014). While most of these studies assess breadth of semantic knowledge (number of words known) it makes sense that depth and richness would increase over time as well. Semantic knowledge is dynamic; retrieving and producing semantic representations across time and contexts changes the nature of those representations and the relations among lexical items (Oppenheim, Dell, & Schwartz, 2007, 2010). We propose the hippocampus, in concert with other neural structures, plays a role in updating semantic representations and in maintaining the relations among items over time.

More specifically, it is our view that the acquisition of semantic knowledge and the updating, or enriching, of existing knowledge with new features are both forms of new learning. It is also our view that these forms of learning both require relational binding processes afforded by the hippocampus. Support for this view comes from the well-established role of the hippocampus in new learning (i.e., the acquisition of novel semantic, conceptual or event knowledge) and the deleterious consequences of hippocampal pathology on new declarative (relational) memory (e.g., Eichenbaum & Cohen, 2001; Squire, 1992). Moreover, recent work points to the role of the hippocampus in the dynamic integration and updating of conceptual and event knowledge (i.e., the acquisition of the new features of a current experience together with shared information from previously stored knowledge) (e.g., Croft et al., 2010; Mack, Love, & Preston, 2016, 2018; Schlichting & Preston, 2015). Thus, the hippocampus plays a critical role in relational learning, whether of entirely new information or in the updating of existing knowledge with new features (Duff et al., in press, for broader review and discussion).

Methodologically, an ideal way to test the proposal that the hippocampus plays a role in updating semantic representations and in maintaining the relations among items over time would be to systematically probe semantic knowledge in patients with hippocampal amnesia from the onset of their memory loss. This type of study, for our group at least, is not possible. Indeed, in our study (Klooster & Duff, 2015) all the participants with amnesia had been living with memory impairment for about 20 years, on average (with a range of 14–34 years). The approach we take here is to study a new group of healthy comparison participants. Critically, in contrast to our previous work where we matched the patients to healthy individuals who were the same age as the patients at the time of testing, here, we recruited healthy individuals who were matched to each patient’s age of onset (AoO) of their brain damage and resulting amnesia. Thus, this new comparison group was the same age as the patients with amnesia were when they incurred their brain damage and amnesia. We recruited 15 new non-brain injured comparison participants, 3 individually matched to each amnesic patient on sex, handedness, and level of education. These healthy participants were native English speakers, free of neurological and psychiatric disease.

Data from these 15 participants were compared with previously published data from 5 amnesic patients with bilateral hippocampal damage and severe impairments in declarative memory, and 15 healthy participants matched to the patients’ current age (CA) (as reported in Klooster & Duff, 2015). Participants completed three tasks; (1) a features-listing task requiring participants to list features of concepts, (2) a senses-listing tests where participants list different senses or different meanings of common words, and (3) The Word Associates Test, a receptive measure of depth of vocabulary knowledge which requires participants to identify word associates from a list of competitors. The procedures were an exact replication of Klooster and Duff (2015).

Several outcomes are possible. First, the healthy younger AoO group may perform similarly to the amnesic patients yet significantly worse than the CA group. This outcome would support our hypothesis that hippocampal relational binding supports the updating and enriching of semantic knowledge over time, and in the more than 20 additional years of normal hippocampal functioning the CA group has had over both the AoO and amnesia groups, has acquired significantly more semantic knowledge about known words. Second, the healthy younger AoO and the healthy older CA comparison groups may perform significantly better than the amnesia group but not significantly different from each other. This outcome would suggest that semantic richness and depth of vocabulary knowledge do not significantly increase over time in the healthy brain (or not as assessed here) and that the observed deficit in remote semantic knowledge in amnesia may be best explained as a loss of knowledge rather than as a failure to update existing semantic knowledge. Finally, the AoO group may perform worse than the CA group, but better than the amnesic group. This result would suggest that the hippocampus may support semantic knowledge by both updating new content into existing representations while also maintaining existing knowledge from attrition.

The critical question here was: can we explain the deficit in remote semantic knowledge in amnesia as a failure to update representations with new semantic information since the onset of their injury and memory deficit? To answer this question, we look to the performance on measures of semantic richness and depth of lexico-semantic knowledge of the age of onset matched healthy comparison participants.

2. Results

2.1. Features

On the features task (Fig. 1), Group (Amn, AoO, CA) significantly predicted the number of features produced (χ2(2) = 13.42, p < 0.005 compared to a model without group), with the AoO group (M = 19.66, SE = 3.05) performing significantly worse than CA group (M = 22.31, SE = 3.05, p < 0.01) and both healthy groups performing better than the patients with amnesia (M = 9.98, SE = 2.67, p < 0.005).

Fig. 1. Features Task.

Fig. 1.

Open in a new tab

Number of features produced by the hippocampal amnesic participants (Amn), younger healthy comparisons matched to the patients’ age of onset (AoO), and older healthy comparisons matched to the patients current age (CA). Group means are represented by the black bars. Participant means are represented with black dots. The violin plots (grey shapes) represent the probability density of all observations (35 per participant) with wider areas indicating a greater number of observations at that value within that group.

Including Familiarity did not improve model fit (p > 0.15). Familiarity did not differ between groups (p > 0.2) and all groups gave a mean familiarity score of more than 8.25/9, indicating that these words were subjectively (and equally) well known.

2.2. Senses

On the senses task (Fig. 2), Group significantly predicted the number of senses produced (χ2(2) = 18.68, p < 0.0001 compared to a model without group), with the AoO group (M = 3.23, SE = 0.23) performing significantly worse than CA group (M = 3.64, SE = 0.23, p < 0.005) and both healthy groups performing better than the patients with amnesia (M = 2.49, SE = 0.22, p < 0.005).

Fig. 2. Senses Task.

Fig. 2.

Open in a new tab

Mean number of senses produced by the hippocampal amnesic participants (Amn), younger healthy comparisons matched to the patients’ age of onset (AoO), and older healthy comparisons matched to the patients current age (CA). Group means are represented by the black bars. Participant means are represented with black dots. The violin plots (grey shapes) represent the probability density of all observations (41 per participant) with wider areas indicating a greater number of observations at that value.

Familiarity differed significantly across groups (p < 0.01), with the AoO group (M = 7.27) rating the target words as significantly less familiar than the CA group (M = 8.86) and the participants with amnesia (M = 8.40). However, including familiarity did not improve model fit (p > 0.19), indicating that differences in familiarity could not explain the differences in senses produced between groups.

2.3. WAt

On the Word Associates Test (Fig. 3), Group (Amn, AoO, CA) significantly predicted the number of associates correctly identified (χ2(2) = 23.01, p < 0.0001 compared to a model without group). The patients (M = 3.11, SE = 0.094) identified significantly fewer correct associates on the WAT than the AoO (M = 3.505, SE = 0.10, p < 0.001) and CA (M = 3.69, SE = 0.10, p < 0.001) groups, while the CA group performed better than AoO (p < 0.001) group.

Fig. 3. Word Associates Test (WAT).

Fig. 3.

Open in a new tab

Mean WAT performance by the hippocampal amnesic participants (Amn), younger healthy comparisons matched to the patients’ age of onset (AoO), and older healthy comparisons matched to the patients current age (CA). Group means are indicated with black bars. Participant means are represented with black dots. The violin plots (grey shapes) represent the probability density of all observations (40 total test items) with wider areas indicating a greater number of observations at that value.

2.3.1. Synonyms

For the synonym sub-scale, group predicted the number of synonyms correctly identified (χ2(2) = 9.99, p < 0.01 compared to a model without group). The CA (M = 1.653, SE = 0.084) group identified significantly more matching synonyms than both the patients with amnesia (M = 1.465, SE = 0.095, p < 0.05) and the AoO (M = 1.547, SE = 0.084, p < 0.05) group. The patients with amnesia and the AoO group did not significantly differ on synonyms (p = 0.19).

2.3.2. Collocates

On the collocate sub-scale, group predicted the number of collocates correctly identified (χ2(2) = 22.24, p < 0.0001 compared to a model without group), patients with amnesia (M = 1.645, SE = 0.106) performed significantly worse than both the AoO (M = 1.958, SE = 0.094, p < 0.001) and CA (M = 2.034, SE = 0.094, p < 0.001) groups, while CA and AoO did not differ (p < 0.124).

3. Discussion

Semantic memory is dynamic and expands across the lifespan. The older CA comparison group performed significantly better than the younger AoO group on all tasks suggesting that with age, healthy individuals add new features, senses, and associates to existing semantic representations. Recall that all the stimuli tested in this study were rated as highly familiar to the participants. With many more years of experience and use, the older CA group had significantly more information associated with each lexical/conceptual item. That vocabulary size (breadth) continues to increase across the lifespan is well documented and stands in stark contrast to the decline observed across other cognitive domains including executive functions, episodic memory abilities, reasoning skills, spatial visualization, and processing speed (Buckner, 2004; Fjell, Sneve, HGrydeland, Storsve, & Walhovd, 2017; Light, 1991; Salthouse, 1993; Salthouse, 2009). The results reported here show that it is not just breadth of vocabulary that increases with age but also depth and richness of vocabulary knowledge. Thus, the previously reported difference in performance between the patients with amnesia and their CA matched comparison participants becomes all the more striking as it highlights the amount of semantic knowledge the patients would have likely acquired over time if not for their hippocampal pathology. It is the shared hippocampal damage that appears to be driving these results across the patients with amnesia of different etiologies and not differences in the extent of medial and lateral temporal lobe damage. The patients with damage limited to the hippocampus performed identically to those with more widespread damage and both sub-groups performed significantly worse than their CA matched comparison participants (Klooster & Duff, 2015).

The main goal of the current study was to test the idea that the observed deficits in remote semantic knowledge in patients with hippocampal amnesia (Klooster & Duff, 2015) were the result of a failure to update existing semantic representations with new information since the onset of their amnesia. We found that a group of healthy comparison participants matched to the age the patients were at the time of onset (AoO) of hippocampal damage and amnesia performed significantly better than the patients with amnesia. That the CA comparison group performed significantly better than the AoO comparison group provides clear evidence for the idea that there is an increase in the amount of information that is associated with lexical items over time. The observation that the patients with amnesia have failed to acquire the amount of semantic knowledge evident in the CA group is broadly consistent with a failure to update highly familiar words with new or additional features over time. A failure to update interpretation also fits with a large literature documenting deficits in new semantic learning in amnesia (e.g., Gabrieli et al., 1988) and with work showing hippocampal contributions to the updating and integration of concept-relevant information during learning (Mack, Love, & Preston, 2016, 2018; Schlichting & Preston, 2015). These data, especially the collocational deficit on the WAT, are also consistent with findings that hippocampal damage is associated with deficits in statistical learning and tracking co-occurrence information from the (linguistic) environment over time (Covington, Brown-Schmidt, & Duff, 2018; Schapiro, Gregory, Landau, McCloskey, & Turk-Browne, 2014). To be clear, we do not claim hippocampal pathology and amnesia abolish the ability to update existing semantic information with any new information. In fact, data suggest that patient H.M. was able to add some new information to his semantic knowledge under some limited circumstances (e.g., O’Kane, Kensinger, & Corkin, 2004). Rather the data here lead us to the conclusion that the individuals with hippocampal pathology have added significantly less information to their existing semantic knowledge stores than would be predicted in the context of normal, hippocampal functioning.

Indeed, the fact that the AoO group performed significantly better than the patients with amnesia raises the possibility that hippocampal damage and memory loss may contribute to reduced updating of semantic knowledge. In the context of theories that argue word learning is a protracted process and never fully complete (Carey, 2010; McMurray et al., 2012), it can be difficult to tease apart, and perhaps more difficult to test in patients, a failure to update new information from the loss of existing knowledge. That said, the current findings make contact with multiple literatures and raise interesting questions that may inform and advance future work in this area.

In contrast to consolidation theories of memory that posit a transfer of knowledge from the hippocampus to neocortex, other theories and proposals suggest that episodic memory always depends on the hippocampal complex, for the life of that memory (Nadel & Moscovitch, 1997; Yonelinas, Ranganath, Ekstrom, & Wiltgen, 2019). Evidence for this proposal comes from deficits in remote episodic memory in patients with unilateral temporal lobe epilepsy and excisions (Viskontas, McAndrews, & Moscovitch, 2000). The current data suggest that remote semantic memory, like remote episodic memory, may also always depend on the hippocampus and that in the absence of a functional hippocampus, we can observe disruptions in semantic memory. Such shared dependency on the hippocampus is consistent with the view that both semantic memory and episodic memory rely critically on the hippocampus (Eichenbaum & Cohen, 2001; Gabrieli et al., 1988; Squire & Zola, 1998). Future work testing the long-term dependency of remote semantic memory on the hippocampus could use measures of semantic memory depth such as those used here. In the memory literature, there has been considerable effort in the development of tools and measures to capture the richness and detail of episodic memory while such efforts have lagged behind for capturing similar complexities in semantic memory.

Models of language learning emphasize the role of one’s use and experience with language in driving change across the lifespan (e.g. Chang, Dell, & Bock, 2006; Ramscar et al., 2014; Hartshorne & Germine, 2015). It is possible that patients with amnesia have degraded remote semantic memory because they have fewer opportunities to use and experience language. Such a scenario might result in fewer occasions to acquire new semantic features of a lexical item through experience (i.e., why patients perform worse than the CA group) or might result in a loss of knowledge (or weakened access) through lack of use (i.e., why patients perform worse than the AoO group). While there are many reports of social isolation in patients with amnesia (Davidson, Drouin, Kwan, Moscovitch, & Rosenbaum, 2012; Tate, 2002), the participants in the current study do not differ from healthy participants in the size of their social networks or report differences in opportunities for social engagement (Beadle & Duff, 2015). That said, we do not know about the linguistic quality or richness of those interactions. The observed deficits in learning new semantic information and in statistical learning in amnesia may suggest that even if the patients had quantitatively and qualitatively similar opportunities for linguistically rich social interaction, their hippocampal damage may prevent them from adding to, tracking, and strengthening existing semantic representations.

An interesting future direction in this line of work with healthy individuals would be to understand the (dis)similarities in rates of semantic and episodic memory accumulation and decline with age. There are competing theories about the (in)dependence of semantic and episodic memory on the structure and function of the hippocampus and related MTL structures. Here we report that the depth of semantic knowledge increases with age, yet it is well established that episodic memory complaints and decline are common in healthy older adults, particularly after the sixth decade (Schaie, 1996) and these changes are linked to declines in hippocampal structure and function (Fjell et al., 2013; Raz, Ghisletta, Rodrigue, Kennedy, & Lindenberger, 2010; Walhovd et al., 2011). We should note that the CA participants were between the ages of 54 and 62 (approaching the age where episodic memory and hippocampal decline are often first observed). Future work that tracks semantic and episodic memory function, and its relation to structural changes of the hippocampus, well beyond the sixth decade (i.e., participants older than those studied here) and using tasks that are sensitive to the richness of semantic depth is warranted.

An intriguing finding was that the AoO group rated target words as significantly less familiar in the Senses task than the CA group. We believe this result may provide additional support for the notion that lexico-semantic knowledge grows over time. While including familiarity in the analysis of the Senses task did not improve model fit (suggesting differences in familiarity could not explain the group difference), it is indeed interesting that these known, common words were rated as less familiar by the younger AoO group than the older CA group. This finding may suggest that ratings of familiarity strengthen over time with increased experience with, and exposure to, a word even when the word is well known to the individual. It is also interesting that this difference in familiarity is evident only in the Senses task, which assesses knowledge of words than can have multiple meanings. We speculate that familiarity for the words in the Senses task may reflect the lower levels of knowledge of, or experience with, the multiple meanings of the words for the younger AoO group relative to the older CA group. These possible interpretations are speculative and warrant further investigation.

There are limitations to the current study. One limitation is the cross-sectional design. A longitudinal study of increasing semantic richness with age, or as we would predict, of decreasing semantic richness in the context of hippocampal pathology (e.g., in mild cognitive impairment or Alzheimer’s disease) is warranted. A second limitation is the lack of detailed psychometric data on the comparison participants. It is possible that there are idiosyncratic factors driving group differences between the CA and AoO comparison groups such as intelligence or executive functions. In designing this study, we made several decisions that we believe may address, or reduce, significant concerns. Regarding intelligence, we closely matched the two comparison groups on education, which has been shown to be highly correlated with IQ and is often used as a proxy for intelligence (Deary & Johnson, 2010). While we do not have IQ data on the comparison participants, education levels of the two comparison groups are nearly identical (CA mean (standard deviation) education in years = 16.1(1.9); AoO = 16.2. (1.9), suggesting that the IQ levels of the two groups are very similar as well. It is also possible that executive function ability contributes to task performance and could differ among comparison groups. However, if executive function ability were driving task performance, we would have predicted that the older CA group (mean age = 58.5) would perform worse than the younger AoO group (mean age = 38.6) given the literature on age related declines in executive function ability (e.g., Buckner, 2004; Fjell et al., 2017). Our findings were the opposite of this prediction. Nevertheless, future studies should replicate these findings with larger, well characterized samples of healthy participants. Finally, the small sample size of the patients is a limitation. Testing three healthy participants matched to each patient increases power to detect group differences and statistical techniques appropriate for smaller samples have been employed but replicating this work in more patients with amnesia would help solidify the results.

3.1. Conclusions

In summary, we found that in addition to growth in vocabulary size with age in healthy individuals, semantic memory also becomes deeper and richer over time as additional information is added to existing representations. A group of healthy participants matched to the patients’ age at the time of onset of hippocampal damage performed significantly better than the patients but worse than the chronologically matched comparison group. These data lead us to the conclusion that a failure to update semantic memory over time may explain (at least some of) the previously observed deficits in amnesia, but that loss of information also likely occurred. Longitudinal data from patients with hippocampal pathology, however, are required to fully substantiate this interpretation and would provide a critical test of our proposal. Taken together with our previous study (Klooster & Duff, 2015), these results suggest that the hippocampus may support semantic memory across the lifespan.

4. Methods and materials

4.1. Tasks

Participants completed the features, senses, and Word Associates Tests as described in Klooster & Duff, 2015.

4.2. Familiarity survey

Participants rated their familiarity with the words tested on a scale of 1 (not at all familiar) to 9 (extremely familiar) prior to the feature and senses-listing tasks. All stimuli were words that entered the lexicon prior to the onset of amnesia in the patient group.

4.3. Statistical methods

Linear mixed effects analyses, conducted in R (version 3.4.2, R Core Team, 2017) using the lme4 package (Bates, Maechler, Bolker, & Walker, 2015), were used to analyze performance on each task. Degrees of freedom and p-values were calculated using the lmerTest package (Kuznetsova, Brockhoff, & Christensen, 2016). The number of correct responses produced (features, senses, or associates) was predicted from the fixed effect of Group (Amn, AoO, CA). Random intercepts for subject and item allowed for examination of the influence of Group while controlling for the influence of different means associated with the subjects and items sampled. We began with a null model for the dependent variable with random effects of subject and item. Predictors were added incrementally to assess model improvement. Chi-square tests on the log-likelihood values were used to assess model fit. Planned group comparisons were performed on estimated marginal means using the emmeans package (Lenth, 2017) and corrected for multiple comparisons using the Holm procedure. The data are displayed using violin plots in an attempt to communicate the full distribution of the observations.

Statement of Significance.

The results document richer semantic memory in older adults than younger healthy participants. The results suggest that the hipocampus may play a role in updating and maintaining semantic memory over time. These findings inform theories of hippocampal contributions to semantic memory and views of semantic richness in healthy aging.

Acknowledgments

Supported by: Grant sponsor: NIDCD; Grant number: R01 DC011755 to MCD.

Appendix

See Table A1.

Table A1.

Participant demographics and patient neuropsychological profile.

Subj. Sex Hd. Cur. Age Ons. Ed. Ety. Damage HC Vol. IQ VIQ Voc. GMI
1846 F R 51 30 14 Anoxia Bilateral HC −4.23 84 88 8 57
1951 M R 62 28 16 HSE Bilateral HC + MTL −8.10 106 107 10 57
2308 M L 58 43 16 HSE Bilateral HC + MTL N/A 98 95 11 45
2363 M R 58 42 18 Anoxia Bilateral HC −2.64 98 112 12 73
2563 M L 59 45 16 Anoxia Bilateral HC N/A 102 91 9 75
Amn. 1F 4M 3R 2L 57.6 (± 4.0) 37.6 (± 8.0) 16.0 (± 1.4) 95.4 98.8 10 59
AoO 3F 12M 12R 3L 36.8 (± 10.1) – 16.1 (± 1.9)
CA 3F 12M 12R 3L 58.5 (± 4.5) – 16.2 (± 1.9)
Open in a new tab

Amn. = amnesic patients with bilateral hippocampal damage. AoO = healthy comparison participants matched to age of onset of hippocampal damage and amnesia for patients. CA = healthy comparison participants matched to the patients’ current age. Hd. = Handedness. Cur. Age = age at testing. Ons = patients’ age of onset of brain damage and resulting amnesia. Ed. = years of completed education. Ety. = Etiology of brain damage. HSE = Herpes Simplex Encephalitis. HC = hippocampus. + MTL = damage extending into the greater medial temporal lobes. N/A = no available data. HC Vol = Volumetric data obtained from high-resolution MRI analyses. Hippocampal z-scores represent the combined (left and right hemisphere) studentized residuals of hippocampal volume relative to a comparison group (see Allen, Tranel, Bruss, & Damasio, 2006 for additional details). IQ = Wechsler Adult Intelligence Scale–III Full-scale Intelligence Quotient. VIQ = Wechsler Adult Intelligence Scale–III Verbal Intelligence Quotient. Voc = Wechsler Adult Intelligence Scale–III Vocabulary subtest age-corrected scaled score. GMI = Wechsler Memory Scale–III General Memory Index. Bolded scores are impaired as defined as 2 or more standard deviations below normative data based on educational and occupational background.

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

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. Bates D, Maechler M, Bolker B, & Walker S (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. 10.18637/jss.v067.i01. [DOI] [Google Scholar]
  3. Beadle J & Duff MC (2015). Impact of damage to hippocampus, amygdala, and ventromedial prefrontal cortex on social network size In Poster presentation at the Cognitive Neuroscience Society Conference, San Francisco, CA. [Google Scholar]
  4. Buckner R (2004). Memory and executive function in aging and AD: Multiple factors that cause decline and reserve factors that compensate. Neuron, 44(1), 195–208. [DOI] [PubMed] [Google Scholar]
  5. Carey S (2010). Beyond fast mapping. Language Learning and Development, 6(3), 184–205. 10.1080/15475441.2010.484379. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chang F, Dell GS, & Bock K (2006). Becoming syntactic. Psychological Review, 113(2), 234 10.1037/0033-295X.113.2.234. [DOI] [PubMed] [Google Scholar]
  7. Cohen NJ, & Eichenbaum H (1993). Memory, amnesia, and the hippocampal system. Cambridge, MA, US: The MIT Press. [Google Scholar]
  8. Covington NV, Brown-Schmidt S, & Duff MC (2018). The Necessity of the Hippocampus for Statistical Learning. Journal of cognitive neuroscience, (Early Access), 1–20. 10.1162/jocn_a_01228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Croft KE, Duff MC, Kovach CK, Anderson SW, Adolphs R, & Tranel D (2010). Detestable or Marvelous? – Neuroanatomical correlates of character judgments. Neuropsychologia, 48, 1789–1801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Davidson PS, Drouin H, Kwan D, Moscovitch M, & Rosenbaum RS (2012). Memory as social glue: Close interpersonal relationships in amnesic patients. Frontiers in Psychology, 3, 531 10.3389/fpsyg.2012.00531. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Deary IJ, & Johnson W (2010). Intelligence and education: Causal perceptions drive analytic processes and therefore conclusions. International Journal of Epidemiology, 39(5), 1362–1369. [DOI] [PubMed] [Google Scholar]
  12. Duff MC, Covington NV, Hilverman C, & Cohen NJ (2019). Semantic memory and the hippocampus: Revisiting, reaffirming, and extending the reach of their critical relationship. Frontiers in Human Neuroscience; in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Eichenbaum H, & Cohen NJ (2001). From conditioning to conscious recollection: Memory systems of the brain (No. 35). Oxford University Press on Demand. [Google Scholar]
  14. Fjell A, Sneve M, HGrydeland H, Storsve A, & Walhovd K (2017). The disconnected brain and executive function decline in aging. Cerebral Cortex, 27(3), 2303–2317. [DOI] [PubMed] [Google Scholar]
  15. Fjell AM, Westlye LT, Grydeland H, Amlien I, Espeseth T, Reinvang I, … Walhovd KB (2013). Critical ages in the life course of the adult brain: Nonlinear subcortical aging. Neurobiology of aging, 34(10), 2239–2247. 10.1016/j.neurobiolaging.2013.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gabrieli JD, Cohen NJ, & Corkin S (1988). The impaired learning of semantic knowledge following bilateral medial temporal-lobe resection. Brain and Cognition, 7(2), 157–177. 10.1016/0278-2626(88)90027-9. [DOI] [PubMed] [Google Scholar]
  17. Hartshorne JK, & Germine LT (2015). When does cognitive functioning peak? The asynchronous rise and fall of different cognitive abilities across the life span. Psychological Science, 26(4), 433–443. 10.1177/0956797614567339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kensinger EA, Ullman MT, & Corkin S (2001). Bilateral medial temporal lobe damage does not affect lexical or grammatical processing: Evidence from amnesic patient HM. Hippocampus, 11(4), 347–360. 10.1002/hipo.1049. [DOI] [PubMed] [Google Scholar]
  19. Klooster NB, & Duff MC (2015). Remote semantic memory is impoverished in hippocampal amnesia. Neuropsychologia, 79, 42–52. 10.1016/j.neuropsychologia.2015.10.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kuznetsova A, Brockhoff P, & Christensen R (2016). lmerTest: Tests in Linear Mixed Effects Models. R package version 2.0–33. https://CRAN.R-project.org/package=lmerTest. [Google Scholar]
  21. Laszlo S, & Federmeier KD (2011). The N400 as a snapshot of interactive processing: Evidence from regression analyses of orthographic neighbor and lexical associate effects. Psychophysiology, 48(2), 176–186. 10.1111/j.1469-8986.2010.01058.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lee JL (2008). Memory reconsolidation mediates the strengthening of memories by additional learning. Nature neuroscience, 11(11), 1264 10.1038/nn.2205. [DOI] [PubMed] [Google Scholar]
  23. Lenth R (2017). emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version, (9), 1 https://CRAN.R-project.org/package=emmeans. [Google Scholar]
  24. Light LL (1991). Memory and aging: Four hypotheses in search of data. Annual Review of Psychology, 42(1), 333–376. 10.1146/annurev.ps.42.020191.002001. [DOI] [PubMed] [Google Scholar]
  25. Mack ML, Love BC, & Preston AR (2016). Dynamic updating of hippocampal object representations reflects new conceptual knowledge Proceedings of the National Academy of Sciences, 113(46), 13203–13208. 10.1073/pnas.1614048113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mack ML, Love BC, & Preston AR (2018). Building concepts one episode at a time: The hippocampus and concept formation. Neuroscience Letters, 680, 31–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Manns JR, Hopkins RO, & Squire LR (2003). Semantic memory and the human hippocampus. Neuron, 38(1), 127–133. 10.1016/S0896-6273(03)00146-6. [DOI] [PubMed] [Google Scholar]
  28. McClelland JL, McNaughton BL, & O’reilly RC (1995). Why there are complementary learning systems in the hippocampus and neocortex: Insights from the successes and failures of connectionist models of learning and memory. Psychological Review, 102(3), 419 10.1037/0033-295X.102.3.419. [DOI] [PubMed] [Google Scholar]
  29. McGregor KK, Newman RM, Reilly RM, & Capone NC (2002). Semantic representation and naming in children with specific language impairment. Journal of Speech, Language, and Hearing Research, 45(5), 998–1014. 10.1044/1092-4388(2002/081). [DOI] [PubMed] [Google Scholar]
  30. McKenzie S, & Eichenbaum H (2011). Consolidation and reconsolidation: Two lives of memories? Neuron, 71(2), 224–233. 10.1016/j.neuron.2011.06.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. McMurray B, Horst JS, & Samuelson LK (2012). Word learning emerges from the interaction of online referent selection and slow associative learning. Psychological Review, 119(4), 831 10.1037/a0029872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Metcalfe J, Casal-Roscum L, Radin A, & Friedman D (2015). On teaching old dogs new tricks. Psychological Science, 26(12), 1833–1842. 10.1177/0956797615597912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Nadel L, & Moscovitch M (1997). Memory consolidation, retrograde amnesia and the hippocampal complex. Current Opinion in Neurobiology, 7(2), 217–227. 10.1016/S0959-4388(97)80010-4. [DOI] [PubMed] [Google Scholar]
  34. O’Kane G, Kensinger EA, & Corkin S (2004). Evidence for semantic learning in profound amnesia: An investigation with patient H.M. Hippocampus, 14, 417–425. [DOI] [PubMed] [Google Scholar]
  35. Oppenheim GM, Dell GS, & Schwartz MF (2007). Cumulative semantic interference as learning. Brain and Language, 103(1–2), 175–176. 10.1016/j.bandl.2007.07.102. [DOI] [Google Scholar]
  36. Oppenheim GM, Dell GS, & Schwartz MF (2010). The dark side of incremental learning: A model of cumulative semantic interference during lexical access in speech production. Cognition, 114(2), 227–252. 10.1016/j.cognition.2009.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. O’Reilly RC, & Rudy JW (2000). Computational principles of learning in the neocortex and hippocampus. Hippocampus, 10(4), 389–397. 10.1002/1098-1063(2000)10:4. [DOI] [PubMed] [Google Scholar]
  38. Postle BR, & Corkin S (1998). Impaired word-stem completion priming but intact perceptual identification priming with novel words: Evidence from the amnesic patient HM. Neuropsychologia, 36(5), 421–440. 10.1016/S0028-3932(97)00155-3. [DOI] [PubMed] [Google Scholar]
  39. Core Team R (2017). R: A language and environment for statistical computing Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/. [Google Scholar]
  40. Ramscar M, Hendrix P, Shaoul C, Milin P, & Baayen H (2014). The myth of cognitive decline: Non-linear dynamics of lifelong learning. Topics in Cognitive Science, 6(1), 5–42. 10.1111/tops.12078. [DOI] [PubMed] [Google Scholar]
  41. Raz N, Ghisletta P, Rodrigue KM, Kennedy KM, & Lindenberger U (2010). Trajectories of brain aging in middle-aged and older adults: Regional and individual differences. Neuroimage, 51(2), 501–511. 10.1016/j.neuroimage.2010.03.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Salthouse TA (1993). Speed mediation of adult age differences in cognition. Developmental Psychology, 29(4), 722 10.1037/0012-1649.29.4.722. [DOI] [Google Scholar]
  43. Salthouse TA (2009). When does age-related cognitive decline begin? Neurobiology of Aging, 30(4), 507–514. 10.1016/j.neurobiolaging.2008.09.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Schaie KW (1996). Intellectual development in adulthood: The Seattle longitudinal study. Cambridge University Press. [Google Scholar]
  45. Schapiro AC, Gregory E, Landau B, McCloskey M, & Turk-Browne NB (2014). The necessity of the medial temporal lobe for statistical learning. Journal of Cognitive Neuroscience, 26(8), 1736–1747. 10.1162/jocn_a_00578. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Schlichting ML, & Preston AR (2015). Memory integration: Neural mechanisms and implications for behavior. Current Opinion in Behavioral Sciences, 1, 1–8. 10.1016/j.cobeha.2014.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Squire LR (1992). Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychological Review, 99, 195–231. [DOI] [PubMed] [Google Scholar]
  48. Squire LR, & Zola SM (1998). Episodic memory, semantic memory, and amnesia. Hippocampus, 8(3), 205–211. 10.1002/(SICI)1098-1063(1998)8:3. [DOI] [PubMed] [Google Scholar]
  49. Staudinger UM, Cornelius SW, & Baltes PB (1989). The aging of intelligence: Potential and limits. The Annals of the American Academy of Political and Social Science, 503(1), 43–59. 10.1177/0002716289503001004. [DOI] [Google Scholar]
  50. Tate R (2002). “Emotional and social consequences of memory disorders (R. Tate),” In Baddeley AD, Kopelman MD, and Wilson B (eds.), The Handbook of Memory Disorders (pp. 785–808). 2nd Ed, West Sussex: Wiley. [Google Scholar]
  51. Tulving E (1972). Episodic and semantic memory. Organization of Memory, 1, 381–403. 10.1017/S0140525X00047257. [DOI] [Google Scholar]
  52. Umanath S, & Marsh EJ (2014). Understanding how prior knowledge influences memory in older adults. Perspectives on Psychological Science, 9(4), 408–426. 10.1177/1745691614535933. [DOI] [PubMed] [Google Scholar]
  53. Verfaellie M, Koseff P, & Alexander MP (2000). Acquisition of novel semantic information in amnesia: Effects of lesion location. Neuropsychologia, 38(4), 484–492. 10.1016/S0028-3932(99)00089-5. [DOI] [PubMed] [Google Scholar]
  54. Viskontas IV, McAndrews MP, & Moscovitch M (2000). Remote episodic memory deficits in patients with unilateral temporal lobe epilepsy and excisions. Journal of Neuroscience, 20(15), 5853–5857. 10.1523/JNEUROSCI.20-15-05853.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Walhovd KB, Westlye LT, Amlien I, Espeseth T, Reinvang I, Raz N, … Dale AM (2011). Consistent neuroanatomical age-related volume differences across multiple samples. Neurobiology of Aging, 32(5), 916–932. 10.1016/j.neurobiolaging.2009.05.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Yonelinas AP, Ranganath C, Ekstrom AD, & Wiltgen BJ (2019). A contextual binding theory of episodic memory: Systems consolidation reconsidered. Nature Reviews Neuroscience, 1–12 10.1038/s41583-019-0150-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zeithamova D, & Preston AR (2010). Flexible memories: Differential roles for medial temporal lobe and prefrontal cortex in cross-episode binding. Journal of Neuroscience, 30(44), 14676–14684. 10.1523/JNEUROSCI.3250-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES