Forgetting Rates on the Recency Portion of a Word List Predict Conversion from Mild Cognitive Impairment to Alzheimer’s Disease

Abstract

Amnestic mild cognitive impairment has a greater risk of progressing to Alzheimer’s disease (AD). Consistent with AD patients’ distinctive deficit in consolidating new memory traces, in a recent study we demonstrated that the forgetting rate on the recency portion of a word list differentiates AD from other forms of dementia. In line with this finding, the aim of this study was to investigate whether increased recency forgetting could be a reliable index for predicting amnestic mild cognitive impairment (MCI) patients’ conversion to AD. For this purpose, we compared accuracy in immediate and delayed recall from different portions of a word list in a group of patients with amnestic MCI who converted (C-MCI) or did not convert (S-MCI) to AD during a three-year follow-up period and in a group of normal controls. The results of the present study show that the forgetting from the recency portion of the list (operationalized as a ratio between immediate and delayed recall) was significantly larger in C-MCI than in S-MCI patients. Consistently, the hierarchical logistic regression analyses demonstrated that the recency ratio is a strong predictor of group membership. Similar to what occurs in full-blown AD patients, the results of our study suggest that the increased forgetting rate from the recency portion of the list in C-MCI patients is due to severely reduced efficiency in converting transitory short-term memory representations into stable long-term memory traces. This is consistent with prominent involvement of neuropathological changes in the cortical areas of the medial-temporal lobes and suggests that the recency ratio is a cognitive marker able to identify MCI patients who have a greater likelihood of progressing to AD.

Keywords

Alzheimer’s disease forgetting rate memory disorders mild cognitive impairment recency effect

INTRODUCTION

Mild cognitive impairment (MCI) is a condition of the elderly characterized by the subjective complaint of cognitive difficulty, confirmed by an informant, which is objectified by impaired performance on standard neuropsychological tests [1]. The cognitive impairment of these patients is not severe enough to interfere significantly with the autonomous management of daily living activities and, therefore, does not fulfill the criteria for a diagnosis of dementia. Petersen et al. [2] proposed classifying patients with MCI into sub-types depending on the specific cognitive domains affected. Amnestic MCI is characterized by a subjective memory complaint, which is confirmed by poor performance on episodic memory tests. By contrast, non-amnestic MCI is a cognitive impairment in which cognitive skills other than memory are compromised (e.g., executive or visuospatial functions). A further classification can be made according to whether or not the disorder affects a single or multiple cognitive domains.

In the past two decades, many investigations have highlighted that MCI can be considered a prodromal stage of Alzheimer’s disease (AD). Indeed, the rate of conversion from the amnestic form of MCI to AD is estimated to reach 10–15% per year, which is significantly higher than that of age-matched individuals in the rest of population [3]. Nevertheless, AD is not an inevitable outcome of MCI. In fact, many studies suggest that a significant proportion of these patients will not proceed to AD and that some of them might even return to a cognitive profile compatible with normal aging [4, 5]. With the aim of identifying early biomarkers of AD in the MCI population, a rich literature has been concerned with highlighting the clinical [6], biological [7], neuroimaging [8], and neuropsychological [9 , 11] characteristics that differentiate converters from non converters to AD following the early appearance of the cognitive deficit.

From a neuropsychological perspective, the search for reliable cognitive/memory markers of the memory deficit that could be indicative of an early form of AD has been prominent. Based on the prevalent localization of early brain damage in AD, the memory deficit of MCI patients who eventually convert to AD is considered the result of the precocious involvement of mesio-temporal lobe (MTL) structures by neuropathological changes [12, 13]. Consistent with the role MTL structures play in episodic long-term memory processes (for reviews, see [14]), the memory loss in these patients is likely the expression of defective consolidation mechanisms interfering with the long-term maintenance of memory traces. Consequently, the memory performance of MCI individuals who will convert to AD should be characterized by particularly defective recall accuracy following a delay [15, 16]. By contrast, the memory deficit in non converter MCI patients is likely dependent on a variety of underlying conditions. On the one hand, these patients could also be in a prodromal stage of AD, but with less severe or less aggressive neuropathological changes, thus resulting in a less pronounced delayed recall deficit. On the other hand, the cognitive deficit in these patients could be subtended by other conditions than prodromal AD, including somatic illness, depression or anxiety symptoms, aging-related memory impairment, and prodromal phases of non-AD dementia syndromes, such as frontotemporal or vascular dementias [17 –20]. These cognitive deficits may interfere with the episodic memory functioning, but an increased performance decay passing from immediate to delayed recall trials is generally not expected in these patients.

Consistent with the distinctive deficit of AD patients in consolidating new memory traces, the forgetting rate passing from immediate to delayed recall trials of lists of unrelated words has been frequently proposed as a distinctive marker of the memory loss in converter with respect to non converter MCI patients [9, 21]. In recent years, several papers have emphasized that the decrement in recall accuracy from the terminal portion of a list of unrelated words could be a particularly sensitive marker of AD related memory loss [22, 23]. It is largely agreed, in fact, that different memory components underlie the immediate recall of words from different portions of supraspan word lists. In particular, the enhanced recall from the initial portion of the list (primacy effect) and the less accurate recall from the middle portion have been attributed to the operations of a stable long-term episodic memory mechanism. Conversely, the enhanced recall of the last few words in the list (recency effect), because they were heard immediately before recall, should be the expression of transient storage in a phonological short-term buffer [24, 25]. The increased forgetting rate from the recency portion of the list in AD patients should be due to two concurring factors. On the one hand, a normally functioning short-term memory buffer is responsible for a normal (and in some cases increased) recency effect in immediate recall trials. On the other hand, severely reduced efficiency in converting the transitory short-term memory representations into stable long-term memory traces is responsible for particularly poor recall of recency items in the delayed trials [26]. It should be noted, however, that the recency effect was originally investigated in experimental paradigms using different word lists at each immediate recall trials [24]. For this reason, we should be cautious in accepting an interpretation of the recency effect in immediate recall as the output of a short-term memory buffer when, as in previous works with AD patients, the same list is presented and immediately recalled more times before the delayed recall trial. Analogously, in the same experimental conditions we should also be prudent in interpreting the increased forgetting from the recency portion of the list as resulting from a failure to converting the transitory short-term memory representations into stable long-term memory traces [23 , 27]. In this case, in fact, it is possible that recency items are (at least in part) subject to a learning effect and partially transferred to the long-term memory store. In this case, accuracy decline observed in delayed recall could reflect the contributions of memory decay of the phonological short-term memory store and of a true forgetting from long-term memory.

Bruno et al. [27] operationalized the phenomenon of increased forgetting from the recency portion of the list by proposing a recency ratio (Rr), calculated by dividing the number of terminal words in a list recalled in an immediate memory task by the number of words recalled in a delayed task, as a sensitive index of memory deficit in persons with AD. In a series of papers, they demonstrated 1) that healthy aged individuals exhibiting a high Rr value in a baseline evaluation show a higher rate of conversion to MCI at successive follow-ups than individuals with low Rr values [28], 2) that, consistent with the AD related glutamatergic abnormalities, depressed elderly individuals with low cerebrospinal fluid levels of glutamate have high Rr values, thus suggesting that a high proportion of these individuals actually are in a prodromal phase of AD [29], and 3) that in individuals with MCI, Rr values correlate significantly with the cerebrospinal fluid levels of Aβ₄₂ [30]. In all of these studies, analogous ratios calculated on immediate and delayed recall accuracy from the primacy and mid-portions of the serial position curve had no predictive role. In the same vein, Turchetta et al. [23] found that high Rr values accurately discriminated between AD patients affected by mild to moderate forms of cognitive impairment and patients affected by other forms of dementia (e.g., frontotemporal, Lewy body, and vascular) with comparable levels of cognitive impairment.

The purpose of the present study was to investigate whether a high Rr could be a reliable index for predicting conversion to AD in patients with amnestic MCI. Indeed, based on the assumption that AD patients present, even if only in a preclinical stage, the typical neuropathological changes at the level of the MTL cortical areas [31, 32], we expected that they would also exhibit enhanced forgetting of words confined to the recency position of a word list. Conversely, we expected that non converter individuals with MCI, whose memory deficit did not primarily depend on MTL damage, would exhibit a substantially normal forgetting rate.

Since the experimental paradigm used in the present study consist in repeated presentations and immediate recall trials of the same list, as noted above it is possible that recency items are (at least partially) learned and transferred to the long-term memory store. In order to get some insight about the nature of the increased forgetting from the recency portion of the list, we analyzed whether the two groups of MCI patients improved their recall accuracy of the terminal list items across the successive trials of immediate recall. Indeed, if terminal items would be recalled exclusively from a short-term memory buffer, then no learning effect should be observed across repeated presentation and recall trials. Conversely, accuracy improvement passing from the first to the fifth recall trial would suggest the contribution of long-term memory processes in immediate recall.

METHODS

Subjects

The experimental sample of the current longitudinal study included a cohort of 76 patients diagnosed with amnestic MCI at the first assessment. All patients had been referred to the Alzheimer’s Disease unit of IRCCS Santa Lucia Foundation of Rome from 2000 to 2016. They were submitted to formal clinical, neuropsychological, behavioral, and functional evaluation and a computed tomography (CT) or a magnetic resonance (MR) scan as part of the diagnostic process. The MCI subjects were classified as “amnestic” according to current clinical criteria [2] if they reported: 1) a complaint of memory decline (reported by the subject and confirmed by an informant); 2) objective memory impairment (revealed by scores below age- and education-adjusted norms on at least one of the standard episodic memory tests included in the neuropsychological battery; see below); 3) normal general cognition, as indicated by Mini-Mental State Examination (MMSE) scores above the normality cut-off (>23.8); 4) normal or only mildly impaired activities of daily living (as confirmed by a total Clinical Dementia Rating scale score less than or equal to 0.5); 5) CT or MR brain imaging negative for focal lesions (minimal diffuse changes or minimal lacunar lesions were allowed); and 6) non fulfillment of the clinical criteria for the diagnosis of dementia [33]. Furthermore, the patients had no history of drug/alcohol abuse or any psychiatric or neurological disease.

MCI subjects were invited to follow-up evaluations 12, 24, and 36 months later in which they were again submitted to the clinical examination and the neuropsychological, behavioral, and functional assessments administered in the screening phase. Overall, 33 MCI patients (43% of the entire sample) converted to AD (C-MCI). In the same period, 43 patients (57%) remained in a stable condition of selective cognitive impairment (n = 36) or normalized their performance (n = 7) (S-MCI).

A group of 34 age- and education-matched healthy individuals were also recruited from a healthcare center to serve as normal controls (NC). Inclusion criteria were: 1) absence of neurological or psychiatric disorders; 2) no history of alcohol or drug abuse; and 3) no signs of cognitive decline, as confirmed by a clinical interview and an MMSE score ranging from 27 to 30.

This study was approved by the ethics committee of the Santa Lucia Foundation in accordance with the Declaration of Helsinki and its later amendments. All participants gave their written informed consent prior to their inclusion in the study.

Neuropsychological assessment

All patients with MCI were administered a neuropsychological battery that included tests for assessing the following cognitive areas: declarative memory (15 word-list immediate and delayed recall [34]; prose immediate and delayed recall [35]); visuo-spatial abilities (Copy of drawings; [34]); logical reasoning (Raven’s Coloured Progressive Matrices; [34]) executive functions (Phonemic verbal fluency [34], Modified Card Sorting Test - Number of Criteria achieved [36]); and general cognitive efficiency (MMSE; [37]). Subjects in the NC group were administered only the 15-word list immediate and delayed recall test.

15 Word-list recall test

Since the focus of the present study was on performance on the word-list test, the administration procedure for this task will be presented in detail. This test consists of five consecutive immediate free recall trials of a 15-item word list, followed by a delayed free recall trial. In the first immediate trial, 15 high frequency, semantically unrelated nouns of concrete objects are presented orally by the examiner at a rate of one every second. Immediately following presentation, subjects are asked to recall as many words as possible regardless of the order of list presentation. There is no time constraint for recall. The same procedure is followed in the next four immediate trials. Fifteen minutes after the fifth immediate recall trial (during which subjects are engaged in visual-spatial non memory tasks), subjects are asked to recall as many words from the list as possible without re-presentation (delayed recall). Since the purpose of the present study was to investigate the forgetting rate passing from immediate to delayed recall, we only analyzed accuracy scores on the fifth immediate and delayed recall trials.

Based on the serial position curve shown by the NC group in the fifth immediate trial, primacy, mid-list, and recency scores were computed for each subject by averaging the number of words they recalled in the serial positions 1–3, 4–11, and 12–15, respectively. In order to make comparable the performance scores from the three portions of the serial position curve (which included a different number of serial positions), raw scores were converted into proportional scores relative to the maximum score potentially obtainable. Finally, following Bruno et al. [27], to estimate the performance decrement for the primacy, mid-list, and recency components of the list by taking into account performance accuracy on the immediate test, we computed three ratios (one for each segment of the list) between immediate and delayed performance by applying a slight correction (numerator + 0.05)/(denominator + 0.1) to avoid data loss from cases in which the denominator was 0.

Statistical analysis

One-way ANOVAs were performed to compare patients’ groups on continuous, personal, clinical, and neuropsychological variables. Chi-square tests were performed to analyze gender distribution among groups.

Preliminary analyses carried out with the Shapiro-Wilk’s W test revealed that the distribution of ratio scores relative to different portions of the curve as well as recall scores on the last immediate and delayed trials departed significantly from normality. For this reason, a non parametric ANOVA for independent samples (Kruskal-Wallis H test) was used to compare these scores across groups. In the case the overall group effect resulted significant, the Mann-Whitney U test was used as a post hoc test to investigate the source of this effect.

To test a possible learning effect of the terminal items of the list, we applied a non parametric ANOVA for dependent samples (Wilcoxon matched pairs Z test) to the number of recency items recalled by C-MCI and S-MCI in the 1st versus 5th immediate trial.

A 3-stage hierarchical logistic regression was run in order to examine the independent power of the primacy, mid-list, and recency ratios in discriminating the group membership. We chose a hierarchical approach because we were interested in how the three portions of the word list correctly classified group membership over and above any influence of demographic variables (age, level of education), severity of the cognitive impairment (MMSE scores) and the standard measures of memory functioning from word-list recall. For this analysis, the dependent variable (represented by the group membership) was binarized (C-MCI versus S-MCI). In stage 1, we entered the demographic variables and MMSE scores as predictors; in stage 2, total recall and delayed primacy were entered as predictors; finally, in stage 3 ratios for the three portions of the word-list were entered as predictors. The change in the Block Chi-square allows evaluating how much discriminating power is added to the model by the addition of another block. The accuracy of the final fitted model was then assessed by means of R2 statistics (Cox-Snell).

RESULTS

Groups characteristics

As shown in Table 1, the NC and patient groups did not differ for age, years of education, or gender distribution. A significant difference in average MMSE scores was found between groups because of the better performance of NC compared to S-MCI (p = 0.03) and of S-MCI compared to C-MCI (p < 0.01). The two groups of MCI patients did not differ on any of the tests of the neuropsychological battery, with the exception of immediate and delayed recall of the Short story, where C-MCI scored significantly worse than S-MCI.

Table 1

Demographic and clinical data of controls and MCI patients as well as performance scores of the two groups of MCI patients on the tests of the screening neuropsychological battery

	Group			p
	NC	S-MCI	C-MCI
Sample size	34	43	33
Age (y)	71.6 (5.2)	71.8 (6.3)	74.3 (5.6)	F = 2.3	0.10
Education (y)	9.1 (4.5)	10.3 (4.6)	10.5 (4.6)	F = 0.8	0.42
Sex (M:F)	19:15	22:21	13:20	χ² = 1.9	0.38
MMSE	28.2 (1.0)	27.5 (1.5)	26.6 (1.6)	F = 11.36	0.01
MCST (criteria)	–	3.7 (2.1)	3.8 (1.7)	F = 0.04	0.85
Phonological fluency	–	27.3 (12.1)	28.1 (9.9)	F = 0.09	0.76
Raven’s matrices	–	24.6 (6.3)	25 (5.5)	F = 0.06	0.81
Copy of drawings	–	8.9 (2.4)	9.1 (1.9)	F = 0.1	0.72
Short story IR	–	4.6 (1.1)	3.6 (1.6)	F = 9.5	0.003
Short story DR	–	4.3 (1.7)	2.3 (2.0)	F = 20.5	<0.001

Short story IR, Short story immediate recall; Short story DR, Short story delayed recall; MCST, Modified Card Sorting Test.

Group comparisons

The average proportion of words recalled by the experimental groups in the fifth immediate and delayed recall trials of the word-list as a function of the different portions of the serial position curve are shown in Fig. 1. The ratio values for the primacy, mid-list and recency portions of the serial position curve as well as detailed scores from the 15 word-list test (primacy, mid-list, and recency of the fifth immediate recall trial as well as Total immediate and Total delayed recall) of the two groups of MCI patients and the NC group are reported in Table 2.

Fig.1

Average proportion of recalled words from different portions of the list (primacy, mid-list and recency) during the fifth immediate (- - -) and delayed (—) recall trials in the various participant groups (C-MCI, S-MCI, and NC).

Table 2

Mean values (and SD) of Word list immediate and delayed recall, average proportion of words recalled from the Primacy, Mid-list, and Recency portions of the fifth immediate recall trial and Primacy, Mid-list and Recency ratios of the MCI groups and NC

	Group			Test	p
	NC	S-MCI	C-MCI
Word list immediate	42.3 (10.2)	28.3 (7.0)	23.1 (5.8)	F = 11.7	0.001
Word list delayed	6.0 (2.5)	4.6 (2.2)	2.2 (1.7)	F = 25.4	<0.001
Primacy V trial	0.6 (0.3)	0.5 (0.3)	0.4 (0.2)	F = 4.1	0.02
Mid-list V trial	0.4 (0.1)	0.3 (0.1)	0.3 (0.1)	F = 4.9	0.009
Recency V trial	0.6 (0.2)	0.5 (0.3)	0.5 (0.2)	F = 0.7	0.46
Primacy ratio	1.4 (1.7)	1.8 (2.2)	2.5 (2.3)	χ²_(2,115) = 2.4	0.30
Mid-list ratio	1.3 (0.7)	1.4 (1.1)	1.9 (1.4)	χ²_(2,115) = 5.9	0.05
Recency ratio	1.9 (1.7)	2.2 (2.1)	4.6 (3.2)	χ²_(2,115) = 13.0	0.001

The groups differed in average recall scores in the last immediate (Kruskal-Wallis χ²(2) = 7.35; p = 0.02) and delayed (χ²(2) = 30.65; p < 0.001) recall trials. The primacy ratio was not different between groups (χ²(2) = 2.39; p = 0.30) but the difference in the mid-list ratio approached significance (χ² (2) = 5.92; p = 0.05). Conversely, the between group difference for the Rr (χ²(2) = 13.04; p = 0.001) was highly significant.

Mann-Whitney U tests performed to qualify the group main effects revealed that in the last immediate trial the NC group (M = 8.1; SD = 2.1) scored better than S-MCI (M = 7.0; SD = 2.2.; p = 0.03), which in turn scored better than C-MCI (M = 6.0; SD = 1.7; p = 0.04). Similarly, in the delayed trial, NC (M = 6.1; SD = 2.5) outperformed the S-MCI patients (M = 4.6; SD = 2.3; p = 0.02) who, in turn, were more accurate than C-MCI participants (M = 2.2; SD = 1.8; p < 0.001). Concerning the recency ratio, the NC and the S-MCI group did not differ each other (p = 0.50) and both performed better than C-MCI (p = 0.002 and p = 0.003, respectively).

The Wilcoxon matched pairs Z test comparing the number of recency items recalled by S-MCI patients in the first and last immediate recall trials revealed an approaching significant difference (Z = 1.88; p = 0.06). The same analysis applied to the recency items recalled by C-MCI did not reveal any difference (Z = 1.33; p = 0.18).

Hierarchical logistic regression analyses

The results of the hierarchical logistic regression analyses are reported in Table 3. The first block of the logistic regression model, which included demographic data and MMSE scores as the predictor variables, resulted in a significant improvement in classification from chance (χ² change = 8.59, p = 0.035), with only the MMSE accounting for a significant portion of the variance (Wald = 4.43, p = 0.04). Adding the total recall and delayed primacy as predictors in block 2 led to a statistically reliable improvement of model fit (χ² change = 13.81, p = 0.001), with only these memory measures contributing significantly to the model (total recall: Wald = 5.64, p = 0.02; delayed primacy: Wald = 5.17, p = 0.02). Adding ratios for the three portions of the word-list as predictors (block 3) to block 2 significantly increased the model fit (χ² change = 14, p = 0.003). Results demonstrated that the strongest predictor of group membership was the Rr (Wald = 8.05; p = 0.005) followed by the total recall score (Wald = 5.67; p = 0.02). Neither the demographic variables, MMSE scores, the delayed recall nor the primacy and mid-list ratios contributed significantly to the prediction.

Table 3

Results of the hierarchical logistic regression analysis assessing the predictability of the demographic variables and MMSE (block 1), total immediate and delayed primacy of word list recall (block 2) and primacy, mid-list, and recency ratios (block 3) in classifying patients as belonging to the C-MCI versus S-MCI group

Block	χ ² change	p	-2LL	R ²	% corr
1	8.59	0.035	94.30	0.108	61.3
2	13.81	0.001	80.48	0.258	70.3
3	14	0.003	66.49	0.384	81.3

In order to detect proper Rr cut-off points for discriminating converter from non converter MCI patients, we found that Rr showed a sensitivity of 61% and a specificity of 81% associated with a cut-off of 3.01 (positive predictive value = 72%; negative predictive value = 73%; overall hit ratio = 72%). A cut-off of 2.02, instead, achieved a sensitivity of 73% and a specificity of 65% (positive predictive value = 62%; negative predictive value = 76%; overall hit ratio = 68%).

DISCUSSION

Consistent with the results of a previous study [23], which demonstrated that enhanced forgetting from the recency portion of a word list (operationalized as Rr) is able to accurately discriminate AD patients from patients affected by other forms of dementia, the aim of the present study was to investigate the effectiveness of the Rr in discriminating MCI patients who will convert to AD from those who will not.

For this purpose, we compared accuracy in immediate and delayed recall from different portions of a word list, administered at the time of study entry, of a group of patients with amnestic MCI who, during a three-year follow-up period, converted (47%) or did not convert (57%) to AD, and in a group of NCs. A qualitative analysis of the serial position curves revealed that in the last immediate trial all participants showed the classical serial position curve, characterized by enhanced recall accuracy from the early and terminal positions and significantly less recall accuracy from the mid-list portion of the list. In delayed trial, instead, although a primacy effect was still evident, the recency effect virtually disappeared. This drop in accuracy was particularly evident in the C-MCI group, which in the immediate trial recalled the terminal items at the same level as the other groups but in the delayed trial recalled the same items even less than the mid-list items. The comparison of immediate/delayed recall ratios from the different portions of the list highlighted the particularly striking difference in recall accuracy disclosed by the C-MCI group on the recency portion of the list. Indeed, although the primacy ratios did not differ between groups and the mid-list ratios only approached significance, a highly significant difference was evident for the Rr, with C-MCI obtaining a higher ratio value than both S-MCI and NC groups, which did not differ each other.

The diagnostic value of Rr could be undervalued based on the fact that patients in the S-MCI group scored on average better than C-MCI patients on memory and general cognitive indexes commonly used in laboratory, as the number of words recalled on both the immediate and delayed trials of the word list task and the MMSE score. Indeed, Rr could be interpreted as just an additional index of discrepant verbal memory performance in the two groups of MCI, unnecessary for diagnostic differentiation. However, the hierarchical logistic regression analysis, which assessed the relative contribution of all these variables in predicting AD conversion, demonstrated the added value of Rr for this purpose. Indeed, in this analysis the Rr resulted the strongest predictor of conversion to AD, over and above the predictive value of other variables. This finding emphasizes the unique role of Rr as a valuable diagnostic marker of preclinical AD, even when, alike this case, the usual laboratory memory and general cognitive indexes highlight unbalanced values in converter and stable MCI.

The results of this study are consistent with previous findings [22 , 30] which highlighted Rr as a specific index of AD related memory dysfunction. Carlesimo et al. [22] demonstrated that enhanced forgetting from the terminal portion of a word-list (despite normal forgetting from the other portions) is also a specific finding in amnesic patients with focal MTL damage, which reinforces the view that increased Rr in AD patients is closely related to the early neuropathological changes affecting these areas. As previously noted, the straightforward interpretation of this phenomenon rests on the assumption that different memory systems are implicated in the immediate and delayed recall of terminal words on a list, with the phonological short-term memory buffer responsible for transitory maintenance and immediate recall and the episodic long-term memory system underlying permanent storage and delayed recall. The increased forgetting from recency positions in MCI patients destined to convert in AD should result from proficient immediate recall related to a normally functioning phonological short-term memory store (underlain by cortical networks in the parietal and frontal lobes, which are substantially spared in these patients) and particularly poor delayed recall due to their severely impaired episodic long-term memory system (underlain by the operations of cortical areas in the MTL that are precociously affected in MCI) [38 –40].

The assumption that the increased forgetting rate from the terminal portion of the list in C-MCI patients is expression of a lack of consolidation of these items in the episodic long-term memory store is corroborated by the finding of a lack of any learning effect passing from the 1st to the 5th immediate recall trial (no performance improvement in the recall of recency items). Conversely, the fact that S-MCI patients partially consolidated these items in long-term memory (and for this reason likely showed a lower forgetting rate) is demonstrated by a marginal improvement of recall accuracy across immediate recall trials. It should be noted, however, that the marginal learning effect passing from the 1st to the 5th immediate recall trial in S-MCI patients, if anything increases, and not decreases, the value of the main result of the study. In fact, in this case, the Rr index of S-MCIs found in the study could be resulted as a little augmented compared to the original score (we could say the pure recency score), because the numerator includes an extra score. In other words, the real difference in the Rr index between C-MCI and S-MCI would be more significant than how much reported in the study. It should be noted, however, that a contribution of long-term memory processes is not the only possible explanation of the increased accuracy in recalling terminal items in the list across immediate recall trials. Other hypotheses could be based on performance improvement with practice, perhaps on short-term efficiency, speed of processing, attention, executive control, especially in that group of patients with the less severe decline, just like the S-MCIs are.

C-MCI patients disclosed a primacy ratio analogous to that of NC and S-MCI groups but an approaching significant increase in the mid-list ratio with respect to the other groups. In view of the fact that immediate recall of mid-list items likely reflects the output of stable memory representations [24, 25], the increased forgetting rate from this portion of the list suggests a genuine enhanced decline from the episodic long-term memory store. Since increased immediate/delayed ratios for pre-recency positions has never been observed in full-blown or prodromal AD populations [23 , 28], the replicability of this finding should be replicated in future investigations.

In conclusion, our results are consistent with the findings of several previous studies which argued that the capacity of retention following a delay in memory representations is a useful index for AD-related episodic memory impairment. The added value of our work lies in the demonstration that the Rr, an index of forgetting for the recency portion of a word list, is a cognitive marker able to identify MCI patients who have a higher likelihood of progression to AD. This marker of increased forgetting may add to other indexes of MTL memory dysfunction (for reviews, see [9, 10]) to improve sensitivity and specificity of the neuropsychological diagnosis of prodromal AD. Further research, including neuroimaging (i.e., amyloid PET) and cerebrospinal fluid biomarkers, might expand the current findings and support speculations regarding the differential pathological involvement of cerebral areas which in this study are based only on indirect neuropsychological evidence.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/19-0509r3).

References

Petersen

, Smith

, Waring

, Ivnik

, Tangalos

, Kokmen

(1999) Mild cognitive impairment: Clinical characterization and outcome. Arch Neurol 56, 303–308.

Petersen

, Caracciolo

, Brayne

, Gauthier

, Jelic

, Fratiglioni

(2014) Mild cognitive impairment: A concept in evolution. J Intern Med 275, 214–228.

Mitchell

, Shiri-Feshki

(2009) Rate of progression of mild cognitive impairment to dementia–meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand 119, 252–265.

Palmer

, Fratiglioni

, Winblad

(2003) What is mild cognitive impairment? Variations in definitions and evolution of nondemented persons with cognitive impairment. Acta Psychiatr Scand 107, 14–20.

Perri

, Carlesimo

, Serra

, Caltagirone

, Early Diagnosis Group of the Italian Interdisciplinary Network on Alzheimer’s Disease (2009) When the amnestic mild cognitive impairment disappears: Characterisation of the memory profile. Cogn Behav Neurol 22, 109–116.

Mauri

, Sinforiani

, Zucchella

, Cuzzoni

, Bono

(2012) Progression to dementia in a population with amnestic mild cognitive impairment: Clinical variables associated with conversion. Funct Neurol 27, 49–54.

Frölich

, Peters

, Lewczuk

, Gruber

, Teipel

, Gertz

, Jahn

, Jessen

, Kurz

, Luckhaus

, Hüll

, Pantel

, Reischies

, Schröder

, Wagner

, Rienhoff

, Wolf

, Bauer

, Schuchhardt

, Heuser

, Rüther

, Henn

, Maier

, Wiltfang

, Kornhuber

(2017) Incremental value of biomarker combinations to predict progression of mild cognitive impairment to Alzheimer’s dementia. Alzheimers Res Ther 9, 84.

Beckett

, Donohue

, Wang

, Aisen

, Harvey

, Saito

, Alzheimer’s Disease Neuroimaging Initiative (2015) The Alzheimer’s Disease Neuroimaging Initiative phase 2: Increasing the length, breadth, and depth of our understanding. Alzheimers Dement 11, 823–831.

Gainotti

, Quaranta

, Vita

, Marra

(2014) Neuropsychological predictors of conversion from mild cognitive impairment to Alzheimer’s disease. J Alzheimers Dis 38, 481–495.

10.

Belleville

, Fouquet

, Hudon

, Zomahoun

HTV

, Croteau

(2017) Neuropsychological measures that predict progression from mild cognitive impairment to Alzheimer’s type dementia in older adults: A systematic review and meta-analysis. Neuropsychol Rev 27, 328–353.

11.

De Simone

, Perri

, Fadda

, Caltagirone

, Carlesimo

(2018) Predicting progression to Alzheimer’s disease in subjects with amnestic mild cognitive impairment using performance on recall and recognition tests. J Neurol 266, 102–111.

12.

Markesbery

(2010) Neuropathologic alterations in mild cognitive impairment: A review. J Alzheimers Dis 19, 221–228.

13.

Mufson

, Binder

, Counts

, DeKosky

, Ginsberg

, Ikonomovic

, Perez

, Scheff

(2012) Mild cognitive impairment: Pathology and mechanisms. Acta Neuropathol 123, 13–30.

14.

Squire

, Genzel

, Wixted

, Morris

(2015) Memory consolidation. Cold Spring Harb Perspect Biol 7, a021766.

15.

De Simone

, Perri

, Fadda

, De Tollis

, Turchetta

, Caltagirone

, Carlesimo

(2017) Different deficit patterns on word lists and short stories predict conversion to Alzheimer’s disease in patients with amnestic mild cognitive impairment. J Neurol 264, 2258–2267.

16.

Gainotti

, Quaranta

, Vita

, Marra

(2014) Neuropsychological predictors of conversion from mild cognitive impairment to Alzheimer’s disease. J Alzheimers Dis 38, 481–495.

17.

Visser

, Verhey

, Ponds

, Cruts

, Van Broeckhoven

, Jolles

(2000) Course of objective memory impairment in non-demented subjects attending a memory clinic and predictors of outcome. Int J Geriatr Psychiatry 15, 363–372.

18.

Dubois

, Feldman

, Jacova

, DeKosky

, Barberger-Gateau

, Cummings

, Delacourte

, Galasko

, Gauthier

, Jicha

, Meguro

, O’Brien

, Pasquier

, Robert

, Rossor

, Salloway

, Stern

, Visser

, Scheltens

(2007) Research criteria for the diagnosis of Alzheimer’s disease: Revising the NINCDS–ADRDA criteria. Lancet Neurol 6, 734–746.

19.

Gauthier

, Reisberg

, Zaudig

, Petersen

, Ritchie

, Broich

, Belleville

, Brodaty

, Bennett

, Chertkow

, Cummings

, de Leon

, Feldman

, Ganguli

, Hampel

, Scheltens

, Tierney

, Whitehouse

, Winblad

(2006) Mild cognitive impairment. Lancet Neurol 367, 1262–1270.

20.

Perri

, Carlesimo

, Serra

, Caltagirone

21.

Lee

, Ritchie

, Yaffe

, Cenzer

, Barnes

(2014) A clinical index to predict progression from mild cognitive impairment to dementia due to Alzheimer’s disease. PLoS One 9, e113535.

22.

Carlesimo

, Sabbadini

, Fadda

, Caltagirone

(1995) Different components in word-list forgetting of pure amnesics, degenerative demented and healthy subjects. Cortex 31, 735–745.

23.

Turchetta

, Perri

, Fadda

, Caruso

, De Simone

, Caltagirone

, Carlesimo

(2018) Forgetting rate on the recency portion of a word list differentiates mild to moderate Alzheimer’s disease from other forms of dementia. J Alzheimers Dis 66, 461–470.

24.

Atkinson

, Shiffrin

(1968) Human memory: A proposed system and its control processes. Psychol Learn Motiv 2, 89–195.

25.

Glanzer

(1972) Storage mechanisms in recall. Psychol Learn Motiv 5, 129–193.

26.

Carlesimo

, Sabbadini

, Fadda

, Caltagirone

(1995) Forgetting from long-term memory in dementia and pure amnesia: Role of task, delay of assessment and aetiology of cerebral damage. Cortex 31, 285–300.

27.

Bruno

, Reichert

, Pomara

(2016) The recency ratio as an index of cognitive performance and decline in elderly individuals. J Clin Exp Neuropsychol 38, 967–973.

28.

Bruno

, Koscik

, Woodard

, Pomara

, Johnson

(2018) The recency ratio as predictor of early MCI. Int Psychogeriatr 30, 1883–1888.

29.

Bruno

, Nierenberg

, Cooper

, Marmar

, Zetterberg

, Blennow

, Hashimoto

, Pomara

(2017) The recency ratio is associated with reduced CSF glutamate in late-life depression. Neurobiol Learn Mem 141, 14–18.

30.

Bruno

, Gleason

, Koscik

, Pomara

, Zetterberg

, Blennow

, Johnson

(2019) The recency ratio is related to CSF amyloid beta 1–42 levels in MCI-AD. Int J Geriatr Psychiatry 34, 415–419.

31.

Markesbery

(2010) Neuropathologic alterations in mild cognitive impairment: A review. J Alzheimers Dis 19, 221–228.

32.

Mufson

, Malek-Ahmadi

, Perez

, Chen

(2016) Braak staging, plaque pathology, and APOE status in elderly persons without cognitive impairment. Neurobiol Aging 37, 147–153.

33.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

34.

Carlesimo

, Caltagirone

, Gainotti

, Fadda

, Gallassi

, Lorusso

, Marfia

, Marra

, Nocentini

, Parnetti

(1996) The mental deterioration battery: Normative data, diagnostic reliability and qualitative analyses of cognitive impairment. Eur Neurol 36, 378–384.

35.

Carlesimo

, Buccione

, Fadda

, Graceffa

, Mauri

, Lorusso

, Bevilacqua

, Caltagirone

(2002) Standardizzazione di due test di memoria per uso clinico: Breve Racconto e Figura di Rey. Nuova Riv Neurol 12, 1–13.

36.

Nocentini

, Di Vincenzo

, Panella

, Pasqualetti

, Caltagirone

(2002) La valutazione delle funzioni esecutive nella pratica neuropsicologica; dal Modified Card Sorting Test al Modified Card Sorting Test-Roma Version. Dati di standardizzazione. Nuova Riv Neurol 12, 13–24.

37.

Measso

, Cavarzeran

, Zappalà

, Lebowitz

, Crook

, Pirozzolo

, Amaducci

, Massari

, Grigoletto

(1993) The mini-mental state examination: Normative study of an Italian random sample. Dev Neuropsychol 9, 77–85.

38.

Markesbery

(2010) Neuropathologic alterations in mild cognitive impairment: A review. J Alzheimers Dis 19, 221–228.

39.

Mufson

, Binder

, Counts

, DeKosky

, deToledo-Morrell

, Ginsberg

, Ikonomovic

, Perez

, Scheff

(2012) Mild cognitive impairment: Pathology and mechanisms. Acta Neuropathol 123, 13–30.

40.

Scarpazza

, De Simone

(2016) Voxel based morphometry: Current perspectives. Neurosci Neuroecon 5, 19–35.