Serial Memory Research Articles

Is Verbal Rehearsal Strategic? An Investigation into Overt Rehearsal of Nameable Pictures in 5- to 10-Year-Old Children

ABSTRACT Though verbal rehearsal is a frequently endorsed strategy for remembering short lists among adults, there is ambiguity around when children deploy it, and what circumstantial factors encourage them to rehearse. We recoded data from a recent multilab replication of a serial picture memory task in which children were observed for evidence of task-related speech or lip movements to extract finer-grained detail about how children spoke during the task. With these data, we aimed to better understand the manner in which children rehearse and the task scenarios which elicit overt rehearsal. Children in several countries from 5 to 10 years old were tasked with remembering 2–5 nameable pictures in serial order across a 15-second delay. Coders categorized children’s speech or lip movements as reflecting fixed rehearsal of the last-presented item only, cumulative rehearsal of all the items presented so far, or some attempt at cumulative rehearsal. We found that most children, regardless of age, did not overtly rehearse at all during presentation of the objects or during the delay period. However, children who sometimes overtly rehearsed recalled longer lists of items than children who did not. Though rare, cumulative rehearsal was most frequently observed for list lengths close to the participant’s demonstrated maximum recall length. Critically, on the trials where overt rehearsal was observed, recall improved. This evidence supports previous suggestions that rehearsal strategy, and possibly also its effectiveness, changes with task difficulty, and raises further questions about how verbal rehearsal affects serial recall.

Serial and strategic memory processes in younger and older adults

ABSTRACT We investigated age-related differences in serial and strategic processing during the encoding and retrieval of high-value words. Younger and older adults were presented with word triads positioned left, center, and right, with one word being more valuable than the others. In Experiment 1, younger adults more effectively recalled the middle, high-value word, demonstrating enhanced strategic memory. Younger adults were more likely to initiate recall with a high-value word whereas older adults were equally likely to initiate recall with a left and high-value word. Additionally, older adults were more likely to recall words in their presented order while younger adults strategically recalled successive high-value words. However, both age groups demonstrated strategic processing in Experiments 2 and 3, even without prior knowledge of the high-value word’s location. Thus, serial and strategic processing may differ based on age and task demands, but strategic processing is preserved in older adults in certain contexts.

The reliability of Hebb repetition learning and its association with language and reading in adolescents with intellectual disabilities

Hebb repetition learning (HRL) refers to neurodevelopmental processes characterised by repeated stimulus exposure without feedback, which result in changes in behaviour and/or responses, e.g., long-term learning of serial order. Here, we investigate effects of HRL on serial order memory. The present research aimed to assess the reliability of new HRL measures and investigate their relationships with language and reading skills (vocabulary, grammar, word reading) in adolescents with intellectual disability (ID). A comparison group of children of similar mental age with typical development (TD) was also assessed. ID and TD groups were tested on HRL tasks, evaluating test-retest and split-half reliability. The relationship between HRL and language and reading was analysed after accounting for the influence of mental age and verbal short-term memory. The HRL tasks displayed moderate test-retest (and split-half) reliability, HRL tasks with different stimuli (verbal, visual) were related, and we identified issues with one method of HRL scoring. The planned regression analyses failed to show relationships between HRL and language/reading skills in both groups when mental age, a very strong predictor, was included. However, further exploratory regression analyses without mental age revealed HRL's predictive capabilities for vocabulary in the ID group and reading in the TD group, results which need further investigation and replication. HRL displays promise as a moderately reliable metric and exhibits varied and interpretable predictive capabilities for language and reading skills across groups.

No position-specific interference from prior lists in cued recognition: A challenge for position coding (and other) theories of serial memory

Position-specific intrusions of items from prior lists are rare but important phenomena that distinguish broad classes of theory in serial memory. They are uniquely predicted by position coding theories, which assume items on all lists are associated with the same set of codes representing their positions. Activating a position code activates items associated with it in current and prior lists in proportion to their distance from the activated position. Thus, prior list intrusions are most likely to come from the coded position. Alternative “item dependent” theories based on associations between items and contexts built from items have difficulty accounting for the position specificity of prior list intrusions. We tested the position coding account with a position-cued recognition task designed to produce prior list interference. Cuing a position should activate a position code, which should activate items in nearby positions in the current and prior lists. We presented lures from the prior list to test for position-specific activation in response time and error rate; lures from nearby positions should interfere more. We found no evidence for such interference in 10 experiments, falsifying the position coding prediction. We ran two serial recall experiments with the same materials and found position-specific prior list intrusions. These results challenge all theories of serial memory: Position coding theories can explain the prior list intrusions in serial recall and but not the absence of prior list interference in cued recognition. Item dependent theories can explain the absence of prior list interference in cued recognition but cannot explain the occurrence of prior list intrusions in serial recall.

Is Recognition of Speech in Noise Related to Memory Disruption Caused by Irrelevant Sound?

Listeners with normal audiometric thresholds show substantial variability in their ability to understand speech in noise (SiN). These individual differences have been reported to be associated with a range of auditory and cognitive abilities. The present study addresses the association between SiN processing and the individual susceptibility of short-term memory to auditory distraction (i.e., the irrelevant sound effect [ISE]). In a sample of 67 young adult participants with normal audiometric thresholds, we measured speech recognition performance in a spatial listening task with two interfering talkers (speech-in-speech identification), audiometric thresholds, binaural sensitivity to the temporal fine structure (interaural phase differences [IPD]), serial memory with and without interfering talkers, and self-reported noise sensitivity. Speech-in-speech processing was not significantly associated with the ISE. The most important predictors of high speech-in-speech recognition performance were a large short-term memory span, low IPD thresholds, bilaterally symmetrical audiometric thresholds, and low individual noise sensitivity. Surprisingly, the susceptibility of short-term memory to irrelevant sound accounted for a substantially smaller amount of variance in speech-in-speech processing than the nondisrupted short-term memory capacity. The data confirm the role of binaural sensitivity to the temporal fine structure, although its association to SiN recognition was weaker than in some previous studies. The inverse association between self-reported noise sensitivity and SiN processing deserves further investigation.

Warning—taboo words ahead! Avoiding attentional capture by spoken taboo distractors

ABSTRACT We examine whether the disruption of serial short-term memory (STM) by spoken taboo distractors is due to attentional diversion and unrelated to the underlying disruptive effect of sound on serial STM more generally, which we have argued is due to order cues arising from the automatic pre-categorical processing of acoustic changes in the sound conflicting with serial–order processing within the memory task (interference-by-process). We test whether the taboo-distractor effect is, unlike effects attributable to interference-by-process, amenable to top-down control. Experiment 1 replicated the taboo-distractor effect and showed that it is not merely a valence effect. However, promoting cognitive control by increasing focal task-load did not attenuate the effect. However, foreknowledge of the distractors did eliminate the taboo-distractor effect while having no effect on disruption by neutral words (Experiment 2). We conclude that the taboo-distractor effect results from a controllable attentional-diversion mechanism distinct from the effect of any acoustically-changing sound.

Vitamin B12 deficiency and cognitive impairment: Is there a distinctive cognitive domain affected? A descriptive study in Buenos Aires

AbstractBackgroundVitamin B12 deficiency is associated with cognitive‐behavioral disorders, which may be potentially reversible. No specific associated cognitive domain has been established and alterations in episodic memory, verbal fluency, attention and executive functions, praxia and visuoconstruction have been described. The objective of this work is to evaluate and describe the cognitive domains affected in patients with cognitive impairment associated with vitamin B12 deficiency.MethodCross‐sectional observational study conducted at the Trinidad Mitre sanatorium of Buenos Aires City. Patients with cognitive impairment and vitamin B12 deficiency were analyzed. Information was collected on the cognitive domain affected (taking as altered value the Z score ≤‐1.5 according to age and educational level, obtained in the neurocognitive evaluations), cognitive profile, magnetic resonance imaging features and mean consultation‐supplementation time.Result25 patients with vitamin B12 deficiency and cognitive complaint were analyzed, with a mean age of 68.5 years. 83% presented multidomain mild cognitive impairment being 60% with subcortical pattern. Memory and attention were the most affected domains in 80% of the sample, followed by executive functions with 52%. The most deficient tests were the direct digit test (68%), serial memory (56%), deferred logical memory (52%) and Trail Making B (52%). Nearly 52% presented mild depressive symptomatology. A 56% of the sample presented with severe B12 deficiency with 32% elevation of plasma homocysteine. Supplementation was performed intramuscularly in 68% of the patients with a mean symptomatology‐supplementation onset time of 10.8 months. 92% of the patients presented subjective improvement after supplementation, 32% had control global test evaluation with improvement in all of them and 50% of this group presented control cognitive evaluation with improvement in previously deficient domains.ConclusionAccording to our findings we conclude that these patients with vitamin B12 deficiency present a multidomain pattern with special affectation of memory, attention and executive functions; of mild degree and potentially reversible. The reversibility of the picture could be related to high doses of B12 and early supplementation, using control cognitive assessments as an objective measure of improvement.ReferencesReynolds E. Lancet Neurol. 2006 Nov;5(11):949‐60. Eastley R. Int J Geriatr Psychiatry. 2000 Mar;15(3):226‐33. Robins Wahlin. Biol Psychol. 2001 Jun;56(3):247‐65.

Evidence of a positive effect of verbal cumulative rehearsal on serial order working memory, as early as 4 years old.

The use of a verbal rehearsal strategy (repeating the items to be remembered to oneself in serial order) has been identified as a key factor in explaining working memory (WM) development. However, the debate remains open with regard to the age at which children are able to use it, and the actual benefits of using such a strategy. Numerous methodological constraints to identify WM strategies limit the interpretation of current findings reported in the literature. Moreover, the sequential nature of the cumulative rehearsal strategy means that its use may have a particular influence on maintaining serial order information. The present study aimed to examine (a) whether children aged 4, 5, and 6 (n = 74, 39 female, n = 61, 29 female, and n = 72, 29 female, respectively) were able to implement an instructed verbal rehearsal strategy and (b) the benefits of using such a strategy in terms of WM performance. Specifically, we invited children to use cumulative rehearsal, naming (simple rehearsal), or no strategy when performing verbal WM tasks (immediate serial recall and item and order serial recognition). Moreover, the specific influence of instructed strategies on item and serial order maintenance was assessed in the different age groups. While the benefits of using the naming strategy were limited, cumulative rehearsal instruction increased WM performance in all age groups, particularly for serial order information. Our results demonstrate that, as early as 4 years old, children were able to implement, and benefit from a short period of cumulative rehearsal instruction. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Noun Sequence Statistics Affect Serial Recall and Order Recognition Memory.

Most theories of verbal working memory recognize that language comprehension and production processes play a role in word memory for familiar sequences, but not for novel lists of nouns. Some language emergent theories propose that language processes can support verbal working memory even for novel sequences. Through corpus analyses, we identify sequences of two nouns that resemble patterns in natural language, even though the sequences are novel. We present 2 experiments demonstrating better recall in college students for these novel sequences over the same words in reverse order. In a third experiment, we demonstrate better recognition of the order of these sequences over a longer time scale. These results suggest verbal working memory and recognition of order over a delay are influenced by language knowledge processes, even for novel memoranda that approximate noun lists typically employed in memory experiments.

Landmark Sequence Learning from Real-World Route Navigation and the Impact of Navigation Aid Visualisation Style.

Primacy and recency features of serial memory are a hallmark of typical memory functions that have been observed for a wide array of tasks. Recently, the ubiquity of this serial position effect has been supported for objects learned during navigation, with canonical serial position functions observed for sequence memory of landmarks that were encountered along a route during a highly controlled virtual navigation task. In the present study, we extended those findings to a real-world navigation task in which participants actively walked a route through a city whilst using a navigation aid featuring either realistic or abstract landmark visualisation styles. Analyses of serial position functions (i.e., absolute sequence knowledge) and sequence lags (i.e., relative sequence knowledge) yielded similar profiles to those observed in a lab based virtual navigation task from previous work and non-spatial list learning studies. There were strong primacy effects for serial position memory in both conditions; recency effects only in the realistic visualisation condition; a non-uniform distribution of item-lags peaking at lag +1; and an overall bias towards positive lags for both visualisation conditions. The findings demonstrate that benchmark serial position memory effects can be observed in uncontrolled, real-world behaviour. In a navigation context, the results support the notion that general memory mechanisms are involved in spatial learning, and that landmark sequence knowledge is a feature of spatial knowledge which is affected by navigation aids.

Serial attention to serial memory: The psychological refractory period in forward and backward cued recall

Guided by the conjecture that memory retrieval is attention turned inward, we examined serial attention in serial memory, combining the psychological refractory period (PRP) procedure from attention research with cued recall of two items from brief six-item lists. We report six experiments showing robust PRP effects in cued recall from memory (1–4) and cued report from perceptual displays (5–6), which suggest that memory retrieval requires the same attentional bottleneck as “retrieval” from perception. There were strong direction effects in each memory experiment. Response time (RT) was shorter and accuracy was higher when the cues occurred in the forward direction (left-to-right, top-to-bottom, first-to-last), replicating differences between forward and backward serial recall. Cue positions had strong effects on RT and accuracy in the memory experiments (1–4). The pattern suggested that subjects find cued items in memory by stepping through the list from the beginning or the end, with a preference for starting at the beginning. The perceptual experiments (5–6) showed weak effects of position that were more consistent with direct access. In all experiments, the distance between the cues in the list (lag) had weak effects, suggesting that subjects searched for each cue from the beginning or end of the list more often than they moved through the list from the first cue to the second. Direction, distance, and lag effects on RT and inter-response interval changed with SOA in a manner that suggested they affect bottleneck or pre-bottleneck processes that create and execute a plan for successive retrievals. We conclude that sequential retrieval from memory and sequential attention to perception engage the same computations and we show how computational models of memory can be interpreted as models of attention focused on memory.

Self-generated strategies in the phonological similarity effect.

Strategy use is an important source of individual differences during immediate serial reconstruction. However, not all strategies are equally suited for all tasks. Therefore, assessing participants' dynamic strategy selection across contexts is an important next step for reliable interpretation of individual differences in short-term memory span - in both experimental and clinical settings. Strategy use during reconstruction of phonologically similar and phonologically distinct word sets was directly assessed using a self-report questionnaire. In two experiments, participants reported consistent use of phonological strategies across word sets; however, participants reported additionally using non-phonological strategies (i.e., mental imagery and sentence generation) when tasked with remembering phonologically similar words. In particular, strategy selection was most impacted when the phonologically similar word set was either the only word set or the first word set participants received. When the phonologically similar lists were presented after a classic list of phonologically distinct words, participants continued using the phonological strategies that had been effective for the distinct lists. Moreover, in both experiments, accuracy of phonologically similar lists was better predicted by use of non-phonological strategies than use of phonological strategies. Specifically, reported use of verbalization or rehearsal did not predict accuracy, but participants who reported regularly using mental imagery and/or sentence generation (typically in conjunction with rehearsal) displayed greater serial memory for similar words. These results do not undermine the general assumptions of the phonological similarity effect, but they do indicate that its interpretation is less straightforward than previously thought.

Verbal and visual serial-order memory in deaf signers and hearing nonsigners: A systematic review and meta-analysis.

Previous research suggests Deaf signers may have different short-term and working memory processes compared with hearing nonsigners due to prolonged auditory deprivation. The direction and magnitude of these reported differences, however, are variable and dependent on memory modality (e.g., visual, verbal), stimulus type, and research design. These discrepancies have made consensus difficult to reach which, in turn, slows progress in areas such as education, medical decision-making, and cognitive sciences. The present systematic review and meta-analysis included 35 studies (N = 1,701 participants) that examined verbal (n = 15), visuospatial (n = 10), or both verbal and visuospatial (n = 10) serial-memory tasks comparing nonimplanted, Deaf signers to hearing nonsigners across the life span. Multivariate meta-analyses indicated a significant, negative effect of deafness on verbal short-term memory (forward recall), g = -1.33, SE = 0.17, p < .001, 95% CI [-1.68, -0.98], and working memory (backward recall), g = -0.66, SE = 0.11, p < .001, 95% CI [-0.89, -0.45], but no significant effect of deafness on visuospatial short-term memory, g = -0.055, SE = 0.17, p = 0.75, 95% CI [-0.39, 0.28]. Visuospatial working memory was not analyzed due to limited power. Population estimates for verbal and visuospatial short-term memory were moderated by age wherein studies with adults demonstrated a stronger hearing advantage than studies with children/adolescents. Quality estimates indicated most studies were of fair quality, with only 38% of studies involving Deaf authors. Findings are discussed in the context of both Deaf equity and models of serial memory.

Design and Performance Analysis of 32 × 32 Memory Array SRAM for Low-Power Applications

Computer memory comprises temporarily or permanently stored data and instructions, which are utilized in electronic digital computers. The opposite of serial access memory is Random Access Memory (RAM), where the memory is accessed immediately for both reading and writing operations. There has been a vast technological improvement, which has led to tremendous information on the amount of complexity that can be designed on a single chip. Small feature sizes, low power requirements, low costs, and great performance have emerged as the essential attributes of any electronic component. Designers have been forced into the sub-micron realm for all these reasons, which places the leakage characteristics front and centre. Many electrical parts, especially digital ones, are made to store data, emphasising the need for memory. The largest factor in the power consumption of SRAM is the leakage current. In this article, a 1 KB memory array was created using CMOS technology and a supply voltage of 0.6 volts employing a 1-bit 6T SRAM cell. We developed this SRAM with a 1-bit, 32- × 1-bit, and 32 × 32 configuration. The array structure was implemented using a 6T SRAM cell with a minimum leakage current of 18.65 pA and an average delay of 19 ns. The array structure was implemented using a 6T SRAM cell with a power consumption of 48.22 μW and 385 μW for read and write operations. The proposed 32 × 32 memory array SRAM performed better than the existing 8T SRAM and 7T SRAM in terms of power consumption for read and write operations. Using the Cadence Virtuoso tool (Version IC6.1.8-64b.500.14) and 22 nm technology, the functionality of a 1 KB SRAM array was verified.

The Active Inference Model of Coherence Therapy.

Coherence Therapy is an empirically derived experiential psychotherapy based on Psychological Constructivism. Symptoms are viewed as necessary output from an implicit model of the world. The therapist curates experiences and directs attention toward discovering the model. Rendered explicit, the model is juxtaposed with contradictory knowledge driving memory re-consolidation with resolution of the symptom. The Bayesian Brain views perception and action as inferential processes. Prior beliefs are combined in a generative model to explain the hidden causes of sensations through a process of Active Inference. Prior beliefs that are poor fits to the real world are suboptimal. Suboptimal priors with optimal inference produce Bayes Optimal Pathology with behavioral symptoms. The Active Inference Model of Coherence Therapy posits that Coherence Therapy is a dyadic act of therapist guided Active Inference that renders the (probable) hidden causes of a client's behavior conscious. The therapist's sustained attention on the goal of inference helps to overcome memory control bias against retrieval of the affectively charged suboptimal prior. Serial experiences cue memory retrieval and re-instantiation of the physiological/affective state that necessitates production of the symptom in a particular context. As this process continues there is a break in modularity with assimilation into broader networks of experience. Typically, the symptom produced by optimal inference with the suboptimal prior is experienced as unnecessary/inappropriate when taken out of the particular context. The implicit construct has been re-represented and rendered consciously accessible, by a more complex but more accurate model in which the symptom is necessary in some contexts but not others. There is an experience of agency and control in symptom creation, accompanied by the spontaneous production of context appropriate behavior. The capacity for inference has been restored. The Active Inference Model of Coherence Therapy provides a framework for Coherence Therapy as a computational process which can serve as the basis for new therapeutic interventions and experimental designs integrating biological, cognitive, behavioral, and environmental factors.

VERBAL SHORT-TERM IMPAIRMENT IN DYSLEXIA

Verbal short-term memory impairment is one of the most consistently associated deficits observed indevelopmental reading disorders such as dyslexia. The mechanisms involved in the short-term retention ofinformation for serial tasks may be shared across domains of short-term memory, such as verbal and visuospatialshort-term memory. Few studies have addressed the nature of verbal short-term impairment, particularly with regardto the possibility of temporary serial storage of information. A short-term memory serial order impairment appearsto occur for the retention of verbal and visuospatial sequence information. Impairment in the serial ordering of shorttermmemory is not a characteristic of every individual dyslexic subject and is not specific to dyslexia. Futurestudies should determine whether verbal short-term serial order memory impairment is a risk factor that, inassociation with a deficit in phonological processing, may lead to dyslexia or whether verbal short-term serial ordermemory impairment reflects related deficits unrelated to dyslexia. The purpose of this article is to consider theimplications for dyslexia of one aspect of short-term memory, memory for serialization

Local detour centrality: a novel local centrality measure for weighted networks

Centrality, in some sense, captures the extent to which a vertex controls the flow of information in a network. Here, we propose Local Detour Centrality as a novel centrality-based betweenness measure that captures the extent to which a vertex shortens paths between neighboring vertices as compared to alternative paths. After presenting our measure, we demonstrate empirically that it differs from other leading central measures, such as betweenness, degree, closeness, and the number of triangles. Through an empirical case study, we provide a possible interpretation for Local Detour Centrality as a measure that captures the extent to which a word is characterized by contextual diversity within a semantic network. We then examine the relationship between our measure and the accessibility to knowledge stored in memory. To do so, we show that words that occur in several different and distinct contexts are significantly more effective in facilitating the retrieval of subsequent words than are words that lack this contextual diversity. Contextually diverse words themselves, however, are not retrieved significantly faster than non-contextually diverse words. These results were obtained for a serial semantic memory task, where the word’s location constitutes a significant mediator in the relationship between the proposed measure and accessibility to knowledge stored in memory.

Task-specific auditory distraction in serial recall and mental arithmetic

Previous studies suggest that task-irrelevant changing-state sound interferes specifically with the processing of serial order information in the focal task (e.g., serial recall from short-term memory), whereas a deviant sound in the auditory background is supposed to divert central attention, thus producing distraction in various types of cognitive tasks. Much of the evidence for this distinction rests on the observed dissociations in auditory distraction between serial and non-serial short-term memory tasks. In this study, both the changing-state effect and the deviation effect were contrasted between serial digit recall and mental arithmetic tasks. In three experiments (two conducted online), changing-state sound was found to disrupt serial recall, but it did not lead to a general decrement in performance in different mental arithmetic tasks. In contrast, a deviant voice in the stream of irrelevant speech sounds did not cause reliable distraction in serial recall and simple addition/subtraction tasks, but it did disrupt a more demanding mental arithmetic task. Specifically, the evaluation of math equations (multiplication and addition/subtraction), which was combined with a pair-associate memory task to increase the task demand, was found to be susceptible to auditory distraction in participants who did not serially rehearse the pair-associates. Together, the results support the assumption that the interference produced by changing-state sound is highly specific to tasks that require serial-order processing, whereas auditory deviants may cause attentional capture primarily in highly demanding cognitive tasks (e.g., mental arithmetic) that cannot be solved through serial rehearsal.

A-196 The Relationship Between Attention and Order Errors in Serial Order Memory

Abstract Objective: To investigate if attentional control (AC) predicts the amount of order errors made within serial memory tasks. We also investigated the relationship between AC and item memory (IM), and the relationship between IM and order memory (OM). Methods: Data were collected from undergraduate students at the University of South Alabama. Participants completed three attentional conflict tasks (Bivalent shape task, Global local task, and Flanker task) and three serial memory tasks (symmetry span, digit span, and reading span). The final sample of 112 participants completed all memory tasks and at least two conflict tasks. Confirmatory factor analysis was used to form an AC latent variable from response time difference scores (congruent versus incongruent trials), and IM and OM variables were formed from item and order errors in the memory tasks. Structural equation modeling was used to test correlations among latent variables. Results: The latent AC factor failed to converge. We ran three structural equation models using each conflict task as a predictor variable for both IM and OM. The relationship between IM and OM was significant in all three models. The only other significant pathway across all three models was the relationship between the Flanker task and IM (β = .278, SE = .105, p = .001). Conclusions: Overall, performance on the three conflict tasks did not directly predict order errors. Flanker task performance did directly predict item errors, thus indirectly predicting the amount of order errors in those same tasks.

Improved prediction of early cognitive impairment in multiple sclerosis combining blood and imaging biomarkers.

Disability in multiple sclerosis is generally classified by sensory and motor symptoms, yet cognitive impairment has been identified as a frequent manifestation already in the early disease stages. Imaging- and more recently blood-based biomarkers have become increasingly important for understanding cognitive decline associated with multiple sclerosis. Thus, we sought to determine the prognostic utility of serum neurofilament light chain levels alone and in combination with MRI markers by examining their ability to predict cognitive impairment in early multiple sclerosis. A comprehensive and detailed assessment of 152 early multiple sclerosis patients (Expanded Disability Status Scale: 1.3 ± 1.2, mean age: 33.0 ± 10.0 years) was performed, which included serum neurofilament light chain measurement, MRI markers (i.e. T2-hyperintense lesion volume and grey matter volume) acquisition and completion of a set of cognitive tests (Symbol Digits Modalities Test, Paced Auditory Serial Addition Test, Verbal Learning and Memory Test) and mood questionnaires (Hospital Anxiety and Depression scale, Fatigue Scale for Motor and Cognitive Functions). Support vector regression, a branch of unsupervised machine learning, was applied to test serum neurofilament light chain and combination models of biomarkers for the prediction of neuropsychological test performance. The support vector regression results were validated in a replication cohort of 101 early multiple sclerosis patients (Expanded Disability Status Scale: 1.1 ± 1.2, mean age: 34.4 ± 10.6 years). Higher serum neurofilament light chain levels were associated with worse Symbol Digits Modalities Test scores after adjusting for age, sex Expanded Disability Status Scale, disease duration and disease-modifying therapy (B = −0.561; SE = 0.192; P = 0.004; 95% CI = −0.940 to −0.182). Besides this association, serum neurofilament light chain levels were not linked to any other cognitive or mood measures (all P-values > 0.05). The tripartite combination of serum neurofilament light chain levels, lesion volume and grey matter volume showed a cross-validated accuracy of 88.7% (90.8% in the replication cohort) in predicting Symbol Digits Modalities Test performance in the support vector regression approach, and outperformed each single biomarker (accuracy range: 68.6–75.6% and 68.9–77.8% in the replication cohort), as well as the dual biomarker combinations (accuracy range: 71.8–82.3% and 72.6–85.6% in the replication cohort). Taken together, early neuro-axonal loss reflects worse information processing speed, the key deficit underlying cognitive dysfunction in multiple sclerosis. Our findings demonstrate that combining blood and imaging measures improves the accuracy of predicting cognitive impairment, highlighting the clinical utility of cross-modal biomarkers in multiple sclerosis.