Letter Pairs Research Articles

Defining letter similarity

In letter identification, confusion probabilities are often used to define letter similarity. The goal of this study is to develop a universal, objective method to quantify letter similarity. Since letter similarity may be associated with the number of similar features shared by the letter pair, we started with identifying 22 potential letter features that appear universally across everyday fonts. A san-serif font, Century Gothic, was used for letter feature analysis. Only lowercase letters were considered. Customized Matlab scripts were used to process letter images and to detect and quantify letter features for each letter, following which similarity score between each pair of letters was calculated. There were 22 sets of similarity scores, one for each feature. To assess the amount of variation in letter confusion probability accounted for by the similarity scores, we performed linear regression analyses (considering only the linear terms). Four confusion matrices measured with different fonts under different testing conditions (Bouma, 1971; Geyer, 1977; McGraw et al., 1994) were examined. Despite the usage of different fonts in calculating similarity scores and obtaining confusion matrices, the best fitting models explained about 30% of the variability of letter confusion across the four confusion matrices. Ten letter features (e.g., number of separated parts, parallelism, and closed area) consistently showed significant contributions to the errors in letter identification. These findings suggested that defining letter similarity in a universal, objective way is feasible. Meeting abstract presented at VSS 2017

A theory-driven model of handshape similarity

Following the Articulatory Model of Handshape (Keane 2014), which mathematically defines handshapes on the basis of joint angles, we propose two methods for calculating phonetic similarity: a contour difference method, which assesses the amount of change between handshapes within a fingerspelled word, and a positional similarity method, which compares similarity between pairs of letters in the same position across two fingerspelled words. Both methods are validated with psycholinguistic evidence based on similarity ratings by deaf signers. The results indicate that the positional similarity method more reliably predicts native signer intuition judgements about handshape similarity. This new similarity metric fills a gap in the literature (the lack of a theory-driven similarity metric) that has been empty since effectively the beginning of sign-language linguistics.

The role of allograph representations in font-invariant letter identification.

The literate brain must contend with countless font variants for any given letter. How does the visual system handle such variability? One proposed solution posits stored structural descriptions of basic letter shapes that are abstract enough to deal with the many possible font variations of each letter. These font-invariant representations, referred to as allographs in this paper, while frequently posited, have seldom been empirically evaluated. The research reported here helps to address this gap with 2 experiments that examine the possible influence of allograph representations on visual letter processing. In these experiments, participants respond to pairs of letters presented in an atypical font in 2 tasks-visual similarity judgments (Experiment 1) and same/different decisions (Experiment 2). By using representational similarity analysis (RSA) in conjunction with linear mixed effect models (LMEM; RSA-LMEM) we show that the similarity structure of the responses to the atypical font is influenced by the predicted similarity structure of allograph representations even after accounting for font-specific visual shape similarity. Similarity due to symbolic (abstract) identity, name, and motor representations of letters are also taken into account providing compelling evidence for the unique influence of allograph representations in these tasks. These results provide support for the role of allograph representations in achieving font-invariant letter identification. (PsycINFO Database Record

Deficits in Letter-Speech Sound Associations but Intact Visual Conflict Processing in Dyslexia: Results from a Novel ERP-Paradigm.

The reading and spelling deficits characteristic of developmental dyslexia (dyslexia) have been related to problems in phonological processing and in learning associations between letters and speech-sounds. Even when children with dyslexia have learned the letters and their corresponding speech sounds, letter-speech sound associations might still be less automatized compared to children with age-adequate literacy skills. In order to examine automaticity in letter-speech sound associations and to overcome some of the disadvantages associated with the frequently used visual-auditory oddball paradigm, we developed a novel electrophysiological letter-speech sound interference paradigm. This letter-speech sound interference paradigm was applied in a group of 9-year-old children with dyslexia (n = 36) and a group of typically developing (TD) children of similar age (n = 37). Participants had to indicate whether two letters look visually the same. In the incongruent condition (e.g., the letter pair A-a) there was a conflict between the visual information and the automatically activated phonological information; although the visual appearance of the two letters is different, they are both associated with the same speech sound. This conflict resulted in slower response times (RTs) in the incongruent than in the congruent (e.g., the letter pair A-e) condition. Furthermore, in the TD control group, the conflict resulted in fast and strong event-related potential (ERP) effects reflected in less negative N1 amplitudes and more positive conflict slow potentials (cSP) in the incongruent than in the congruent condition. However, the dyslexic group did not show any conflict-related ERP effects, implying that letter-speech sound associations are less automatized in this group. Furthermore, we examined general visual conflict processing in a control visual interference task, using false fonts. The conflict in this experiment was based purely on the visual similarity of the presented objects. Visual conflict resulted in slower RTs, less negative N2 amplitudes and more positive cSP in both groups. Thus, on a general, basic level, visual conflict processing does not seem to be affected in children with dyslexia.

Domain-specific interference between storage and processing in complex span is driven by cognitive and motor operations.

The role of domain-specific interference in the complex-span paradigm is still controversial. Here we distinguish two operations within the processing task of this paradigm-a cognitive operation to solve the task and a motor operation to give the answer. Their domain-specific interference with maintenance was investigated within the complex-span paradigm. Presentation of the memoranda-six words or four spatially distributed dots-was interleaved with the presentation of letter pairs to be processed. Cognitive operation domain was manipulated by asking participants to make either a rhyme judgment or a symmetry judgment for each letter pair. Motor operation domain was varied by requiring a response by either speaking or pointing. Cognitive load was held constant. Cognitive and motor operation showed separate domain-specific effects on memory: Visuospatial memory was impaired more by visuospatial than verbal cognitive operations and more by manual-spatial than oral motor operations. In contrast, verbal memory performance was worse when verbal cognitive operations and oral motor operations were executed. The results are discussed in relation to the multicomponent model, the time-based resource-sharing (TBRS) model, and the Serial Order in a Box-Complex-Span (SOB-CS) model. They imply that models of working memory must incorporate mechanisms that allow for domain-specific interference in the verbal and the visuospatial domain.

Partial least squares correspondence analysis: A framework to simultaneously analyze behavioral and genetic data.

For nearly a century, detecting the genetic contributions to cognitive and behavioral phenomena has been a core interest for psychological research. Recently, this interest has been reinvigorated by the availability of genotyping technologies (e.g., microarrays) that provide new genetic data, such as single nucleotide polymorphisms (SNPs). These SNPs-which represent pairs of nucleotide letters (e.g., AA, AG, or GG) found at specific positions on human chromosomes-are best considered as categorical variables, but this coding scheme can make difficult the multivariate analysis of their relationships with behavioral measurements, because most multivariate techniques developed for the analysis between sets of variables are designed for quantitative variables. To palliate this problem, we present a generalization of partial least squares-a technique used to extract the information common to 2 different data tables measured on the same observations-called partial least squares correspondence analysis-that is specifically tailored for the analysis of categorical and mixed ("heterogeneous") data types. Here, we formally define and illustrate-in a tutorial format-how partial least squares correspondence analysis extends to various types of data and design problems that are particularly relevant for psychological research that include genetic data. We illustrate partial least squares correspondence analysis with genetic, behavioral, and neuroimaging data from the Alzheimer's Disease Neuroimaging Initiative. R code is available on the Comprehensive R Archive Network and via the authors' websites. (PsycINFO Database Record

Orthographic processing in pigeons (Columba livia)

Learning to read involves the acquisition of letter-sound relationships (i.e., decoding skills) and the ability to visually recognize words (i.e., orthographic knowledge). Although decoding skills are clearly human-unique, given they are seated in language, recent research and theory suggest that orthographic processing may derive from the exaptation or recycling of visual circuits that evolved to recognize everyday objects and shapes in our natural environment. An open question is whether orthographic processing is limited to visual circuits that are similar to our own or a product of plasticity common to many vertebrate visual systems. Here we show that pigeons, organisms that separated from humans more than 300 million y ago, process words orthographically. Specifically, we demonstrate that pigeons trained to discriminate words from nonwords picked up on the orthographic properties that define words and used this knowledge to identify words they had never seen before. In addition, the pigeons were sensitive to the bigram frequencies of words (i.e., the common co-occurrence of certain letter pairs), the edit distance between nonwords and words, and the internal structure of words. Our findings demonstrate that visual systems organizationally distinct from the primate visual system can also be exapted or recycled to process the visual word form.

Unavoidable sets and circular splicing languages

Circular splicing systems are a formal model of a generative mechanism of circular words, inspired by a recombinant behaviour of circular DNA. They are defined by a finite alphabet A, an initial set I of circular words, and a set R of rules. In this paper, we focus on the still unknown relations between regular languages and circular splicing systems with a finite initial set and a finite set R of rules represented by a pair of letters ((1,3)-CSSH systems). When R=A×A, it is known that the set of all words corresponding to the splicing language belongs to the class of pure unitary languages, introduced by Ehrenfeucht, Haussler, Rozenberg in 1983. They also provided a characterization of the regular pure unitary languages, based on the notions of unavoidable sets and well quasi-orders. We partially extend these notions and their results in the more general framework of the (1,3)-CSSH systems.

The effects of alphabet and expertise on letter perception.

Long-standing questions in human perception concern the nature of the visual features that underlie letter recognition and the extent to which the visual processing of letters is affected by differences in alphabets and levels of viewer expertise. We examined these issues in a novel approach using a same-different judgment task on pairs of letters from the Arabic alphabet with 2 participant groups: 1 with no prior exposure to Arabic and 1 with reading proficiency. Hierarchical clustering and linear mixed-effects modeling of reaction times and accuracy provide evidence that both the specific characteristics of the alphabet and observers' previous experience with it affect how letters are perceived and visually processed. The findings of this research further our understanding of the multiple factors that affect letter perception and support the view of a visual system that dynamically adjusts its weighting of visual features as expert readers come to more efficiently and effectively discriminate the letters of the specific alphabet they are viewing. (PsycINFO Database Record

Interaction between object-based attention and pertinence values shapes the attentional priority map of a multielement display.

Previous studies have shown that the perceptual organization of the visual scene constrains the deployment of attention. Here we investigated how the organization of multiple elements into larger configurations alters their attentional weight, depending on the “pertinence” or behavioral importance of the elements’ features. We assessed object-based effects on distinct aspects of the attentional priority map: top-down control, reflecting the tendency to encode targets rather than distracters, and the spatial distribution of attention weights across the visual scene, reflecting the tendency to report elements belonging to the same rather than different objects. In 2 experiments participants had to report the letters in briefly presented displays containing 8 letters and digits, in which pairs of characters could be connected with a line. Quantitative estimates of top-down control were obtained using Bundesen’s Theory of Visual Attention (1990). The spatial distribution of attention weights was assessed using the “paired response index” (PRI), indicating responses for within-object pairs of letters. In Experiment 1, grouping along the task-relevant dimension (targets with targets and distracters with distracters) increased top-down control and enhanced the PRI; in contrast, task-irrelevant grouping (targets with distracters) did not affect performance. In Experiment 2, we disentangled the effect of target-target and distracter-distracter grouping: Pairwise grouping of distracters enhanced top-down control whereas pairwise grouping of targets changed the PRI. We conclude that object-based perceptual representations interact with pertinence values (of the elements’ features and location) in the computation of attention weights, thereby creating a widespread pattern of attentional facilitation across the visual scene.

Pre- and postnatal dietary protein deficiency influences anxiety, memory and social behaviour in the African striped mouse Rhabdomys dilectus chakae

Dietary protein deficiency influences the behavioural phenotypes of mammals. We studied whether protein deficiency during gestation and/or post-weaning heightened anxiety, reduced memory recall and influenced competitive ability in the African striped mouse Rhabdomys dilectus chakae. Mice were subjected to five protein diet treatments, which they received continuously, or were raised on one diet to weaning and switched to an alternate diet post-weaning (Day 16): 1) HP-HP: high protein (24%); first letter pair indicates maternal diet and the second pair indicates offspring diet post-weaning; 2) BP-BP: baseline protein (19%); 3) LP-LP: low protein (10%); 4) HP-LP: switched from high to low protein diet; and 5) LP-HP: switched from low protein to high protein diet. From Day 70, when mice were sexually mature, 20 individuals (10 males, 10 females) per treatment were subjected to three successive experiments, in which we tested their anxiety responses in: 1) an open field arena (time spent in the centre of the open field); 2) novel object recognition (time spent exploring a novel object); and 3) social interactions (excluding BP-BP) in age-matched same-sex dyadic encounters (aggressive, amicable and avoidance behaviours). LP-LP and LP-HP treatment mice spent the least amount of time in the centre of the open field, did not demonstrate object preference compared to the other treatments, and were the most aggressive in dyadic encounters. Our study shows that the systemic effects of protein-deficient diets during early life shapes the behavioural phenotype in R. d. chakae, possibly through early organisation of neuro-biological pathways or competition among littermates.

ERPs Reveal the Time-Course of Aberrant Visual-Phonological Binding in Developmental Dyslexia.

New evidence is accumulating for a deficit in binding visual-orthographic information with the corresponding phonological code in developmental dyslexia. Here, we identify the mechanisms underpinning this deficit using event-related brain potentials (ERPs) in dyslexic and control adult readers performing a letter-matching task. In each trial, a printed letter was presented synchronously with an auditory letter name. Incongruent (mismatched), frequent trials were interleaved with congruent (matched) infrequent target pairs, which participants were asked to report by pressing a button. In critical trials, incongruent letter pairs were mismatched but confusable in terms of their visual or phonological features. Typical readers showed early detection of deviant trials, indicated by larger modulation in the range of the phonological mismatch negativity (PMN) compared with standard trials. This was followed by stronger modulation of the P3b wave for visually confusable deviants and an increased lateralized readiness potential (LRP) for phonological deviants, compared with standards. In contrast, dyslexic readers showed reduced sensitivity to deviancy in the PMN range. Responses to deviants in the P3b range indicated normal letter recognition processes, but the LRP calculation revealed a specific impairment for visual-orthographic information during response selection in dyslexia. In a follow-up experiment using an analogous non-lexical task in the same participants, we found no reading-group differences, indicating a degree of specificity to over-learnt visual-phonological binding. Our findings indicate early insensitivity to visual-phonological binding in developmental dyslexia, coupled with difficulty selecting the correct orthographic code.

Recompression

In this article, we present an application of a simple technique of local recompression, previously developed by the author in the context algorithms for compressed strings [Jeż 2014a, 2015b, 2015a], to word equations. The technique is based on local modification of variables (replacing X by aX or Xa ) and iterative replacement of pairs of letters occurring in the equation by a “fresh” letter, which can be seen as a bottom-up compression of the solution of the given word equation, or, to be more specific, building a Straight-Line Programme for the solution of the word equation. Using this technique, we give new, independent, and self-contained proofs of many known results for word equations. To be more specific, the presented (nondeterministic) algorithm runs in O ( n log n space and in time polynomial in n and log N , where n is the size of the input equation and N the size of the length-minimal solution of the word equation. Furthermore, for O (1) variables, the bound on the space consumption is in fact linear, that is, O ( m ), where m is the size of the space used by the input. This yields that for each k the set of satisfiable word equations with k variables is context sensitive. The presented algorithm can be easily generalised to a generator of all solutions of the given word equation (without increasing the space usage). Furthermore, a further analysis of the algorithm yields an independent proof of doubly exponential upper bound on the size of the length-minimal solution. The presented algorithm does not use exponential bound on the exponent of periodicity. Conversely, the analysis of the algorithm yields an independent proof of the exponential bound on exponent of periodicity.

Cicero’s Textual Relations: The Gendered Circulation of De finibus

I. How Do You Solve a Problem like Caerellia? In a pair of letters to Atticus from late June and early July 45 BCE, Cicero complains about what is usually considered to be a decidedly modern problem: the unauthorized circulation of media on a peer-to-peer network. At the start of the first of these letters to Atticus, whom he blames for this 'leak,' Cicero writes: die mihi, placetne tibi primum edere iniussu meo? hoc ne quidem faciebat, is qui Platonis libros solitus est divulgare, ex quo '[TEXT NOT REPRODUCIBLE IN ASCII].' quid illud? rectumne existimas cuiquam Bruto, cui te auctore [TEXT NOT REPRODUCIBLE IN ASCII]? scripsit enim Balbus ad me se a te quintum de finibus librum descripsisse; in quo non sane multa mutavi, sed tamen quaedam. tu autem commode feceris si reliquos continueris, ne et [TEXT NOT REPRODUCIBLE IN ASCII] habeat Balbus et '[TEXT NOT REPRODUCIBLE IN ASCII] Brutus, sed haec hactenus, ne videar [TEXT NOT REPRODUCIBLE IN ASCII]. etsi nunc quidem maxima mihi sunt haec. ... quo modo autem fugit me tibi dicere? mirifice Caerellia studio videlicet philosophiae flagrans describit a tuis: istos ipsos de finibus habet. ego autem tibi confirmo (possum falli ut homo) a meis earn non habere; numquam enim ab oculis meis afuerunt. tantum porro aberat ut binos scriberent, vix singulos confecerunt. tuorum tamen ego nullum delictum arbitrar itemque te volo existimare; a me enim praetermissum est ut dicerem me eos exire nondum velle. hui, quam diu de nugis! de re enim nihil habeo quod loquar. (Att. 13.21a.1-2 = 327 SB) Tell me, in the first place, do you think it right to publish without my consent? Not even used to do that, a man who was accustomed to circulate Plato's books, from where the phrase comes, Hermodorus trades in tracts. What of this: Do you think it appropriate to give a copy of this text to anyone before Brutus, to whom I dedicated it at your suggestion? Balbus wrote me that he had obtained from you a copy of Book 5 of On Divine Ends. In this book, I didn't make many changes, to be sure, but nevertheless a few. You, however, will do well if you keep a lid on the other books, lest Balbus has an unrevised copy and Brutus a stale one. But enough about these affairs, lest I seem to make a big deal of small things, although they do appear all-important to me at this moment ... How did it escape me to tell you? Caerellia, evidently inflamed wondrously by her passion for philosophy, makes copies from yours; she has On Moral Ends in its entirety. I guarantee this to you--I am human, however, and can be mistaken-that she does not have them from mine. For never were they absent from my eyes; so far from my men having made two copies, they scarcely completed one copy of each book. Nevertheless, I do not think that any wrong was committed by your men and I wish you to think likewise; for I neglected to say that I did not wish these to go public yet. Dear me, how long about nonsense! For I have nothing to say about business. (1) De finibus was, at this time, Cicero's latest philosophical treatise, a five-book examination of the moral philosophies, that is, the ethical ends, of the various ancient schools. (2) It remains an important source for the ancient discourse on ethics. (3) As for the two people who acquired copies of this treatise before Cicero could present Brutus, the dedicatee, with a copy, our knowledge varies considerably. L. Cornelius Balbus is well known to historians: he was a provincial from Gades in Spain who gained Roman citizenship and rose to great heights in Rome, eventually becoming the first naturalized civis to attain the consulship, which he did in 40 BCE. (4) At this point in time, Balbus was, in addition to being Caesar's chief financial agent along with Oppius, one of the dictator's most important counselors, even if his actual role is somewhat opaque. …

A really simple approximation of smallest grammar

In this paper we present a really simple linear-time algorithm constructing a context-free grammar of size 4glog3/2⁡(N/g) for the input string, where N is the size of the input string and g the size of the optimal grammar generating this string. The algorithm works for arbitrary size alphabets, but the running time is linear assuming that the alphabet Σ of the input string can be identified with numbers from {1,…,Nc} for some constant c. Algorithms with such an approximation guarantee and running time are known, however all of them were non-trivial and their analyses were involved. The here presented algorithm computes the LZ77 factorisation and transforms it in phases to a grammar. In each phase it maintains an LZ77-like factorisation of the word with at most ℓ factors as well as additional O(ℓ) letters, where ℓ was the size of the original LZ77 factorisation. In one phase in a greedy way (by a left-to-right sweep and a help of the factorisation) we choose a set of pairs of consecutive letters to be replaced with new symbols, i.e. nonterminals of the constructed grammar. We choose at least 2/3 of the letters in the word and there are O(ℓ) many different pairs among them. Hence there are O(log⁡N) phases, each of them introduces O(ℓ) nonterminals to a grammar. A more precise analysis yields a bound ℓ+4ℓlog⁡(N/ℓ). As ℓ≤g, this yields the desired bound g+4glog⁡(N/g).

Does level of processing affect the transition from unconscious to conscious perception?

Recently, Windey, Gevers, and Cleeremans (2013) proposed a level of processing (LoP) hypothesis claiming that the transition from unconscious to conscious perception is influenced by the level of processing imposed by task requirements. Here, we carried out two experiments to test the LoP hypothesis. In both, participants were asked to classify briefly presented pairs of letters as same or different, based either on the letters' physical features (a low-level task), or on a semantic rule (a high-level task). Stimulus awareness was measured by means of the four-point Perceptual Awareness Scale (PAS). The results showed that low or moderate stimulus visibility was reported more frequently in the low-level task than in the high-level task, suggesting that the transition from unconscious to conscious perception is more gradual in the former than in the latter. Therefore, although alternative interpretations remain possible, the results of the present study fully support the LoP hypothesis.

Functional MRI in medulloblastoma survivors supports prophylactic reading intervention during tumor treatment.

Development of reading skills is vulnerable to disruption in children treated for brain tumors. Interventions, remedial and prophylactic, are needed to mitigate reading and other learning difficulties faced by survivors. A functional magnetic resonance imaging (fMRI) study was conducted to investigate long-term effects of a prophylactic reading intervention administered during radiation therapy in children treated for medulloblastoma. The fMRI study included 19 reading-intervention (age 11.7 ± 0.6 years) and 21 standard-of-care (age 12.1 ± 0.6 years) medulloblastoma survivors, and 21 typically developing children (age 12.3 ± 0.6 years). The survivors were 2.5 [1.2, 5.4] years after completion of tumor therapies and reading-intervention survivors were 2.9 [1.6, 5.9] years after intervention. Five fMRI tasks (Rapid Automatized Naming, Continuous Performance Test using faces and letters, orthographic and phonological processing of letter pairs, implicit word reading, and story reading) were used to probe reading-related neural activation. Woodcock-Johnson Reading Fluency, Word Attack, and Sound Awareness subtests were used to evaluate reading abilities. At the time of fMRI, Sound Awareness scores were significantly higher in the reading-intervention group than in the standard-of-care group (p = 0.046). Brain activation during the fMRI tasks was detected in left inferior frontal, temporal, ventral occipitotemporal, and subcortical regions, and differed among the groups (p < 0.05, FWE). The pattern of group activation differences, across brain areas and tasks, was a normative trend in the reading-intervention group. Standardized reading scores and patterns of brain activation provide evidence of long-term effects of prophylactic reading intervention in children treated for medulloblastoma.

Investigating the Discriminative Power of Keystroke Sound

The goal of this paper is to determine whether keystroke sound can be used to recognize a user. In this regard, we analyze the discriminative power of keystroke sound in the context of a continuous user authentication application. Motivated by the concept of digraphs used in modeling keystroke dynamics, a virtual alphabet is first learned from keystroke sound segments. Next, the digraph latency within the pairs of virtual letters, along with other statistical features, is used to generate match scores. The resultant scores are indicative of the similarities between two sound streams, and are fused to make a final authentication decision. Experiments on both static text-based and free text-based authentications on a database of 50 subjects demonstrate the potential as well as the limitations of keystroke sound.

Discrimination of the Second Generation: Evidence from a Field Experiment in Norway

A major question in labour market research is the extent to which discrimination in employments causes the disadvantages experienced by children of immigrants. This article contributes to the debate by utilising a correspondence test study in which pairs of equivalent resumes and cover letters—one with a Pakistani name and one with a Norwegian name—were sent in response to 900 job openings in the greater Oslo area. The results show that applicants with Norwegian names on average are 25 % more likely to receive a call back for a job interview than equally qualified applicants with Pakistani names. More refined analyses demonstrate that the effect of ethnic background on employment probabilities is larger among men than women and larger in the private sector than in the public sector, and important variations among the occupations included in the study are revealed. In an effort to separate the potentially conflating effects of gender and sector, all applications to gender-segregated occupations were removed from the analyses. Interestingly, the gender differences disappear when exclusively analysing discrimination in gender-integrated occupations by sector. In gender-integrated occupations in the private sector, the gender difference in fact is reversed, indicating that women with minority background are treated less favourably than are minority men in the private sector. These results suggest that the intersection of gender, ethnicity, and sector should be scrutinised more carefully in future field experiments.

Multisensory Integration in Complete Unawareness: Evidence From Audiovisual Congruency Priming

Multisensory integration is thought to require conscious perception. Although previous studies have shown that an invisible stimulus could be integrated with an audible one, none have demonstrated integration of two subliminal stimuli of different modalities. Here, pairs of identical or different audiovisual target letters (the sound /b/ with the written letter “b” or “m,” respectively) were preceded by pairs of masked identical or different audiovisual prime digits (the sound /6/ with the written digit “6” or “8,” respectively). In three experiments, awareness of the audiovisual digit primes was manipulated, such that participants were either unaware of the visual digit, the auditory digit, or both. Priming of the semantic relations between the auditory and visual digits was found in all experiments. Moreover, a further experiment showed that unconscious multisensory integration was not obtained when participants did not undergo prior conscious training of the task. This suggests that following conscious learning, unconscious processing suffices for multisensory integration.