Articulatory loop in ataxic dysarthria

J. Neurolinguistics. Volume 7. Number 112. pp. 115-131, 1992 Printed in Great Britain Articulatory in Ataxic Loop 0911-6044/92 ss.oo+.oo PcrgamonPress Ud zyxwvutsr Dysarthria Laura Chiacchio Luigi Trojan0 Fondazione Clinica de1 Lmoro, Gzmpoli Clinica Neurologica, Seconda Facolta ’ di Medicina, Naples zyxwvut Agostino Cusati AlessandroFilla Clinica Neurologica, Seconda Facolta ’ di Medicina, Naples Clinica Neurologica, Seconda Facolta ’ di Medicina, Naples Dario Grossi Clinica Neurologica, Seconda Facolta ’ di Medicina, Naples ABSMACT The aim of the present paper was to investigate articulatory loop in patients with pure motor speech disturbances. A homogeneous group of 18 patients affected by ataxic dysarthria and without neuropsychological and neuroimaging evidence of cortical damage was s&ctcd. The study comprised four experiments which assessed: (i) immediate and delayed recall of word sequences with and without articulatory suppression; (ii) phonological similarity and word length effects on immediate recall; (iii) articulation rate; and (iv) verbal and spatial span. In agreement with previous studies on patients affected by different types of dysarthria, articulatory loop was found hmctional. However, patients had a slower atticulation rate, and it was argued that this finding can represent a “capacity limitation” of articulatory rehearsal, contributing to the patients’ slight verbal short-term memory defect, A widely accepted mode1 of immediate memory is based on the working memory theory (Badde19 and Hitch 1974; Baddeley 1986). The mode1 foresaw that verbal short-term memory relied upon a language-related slave system, the so-called articulatory loop. Successively, it has been proposed that the articulatory loop is divided into two subcomponents (Baddeley et al. 1984; Vallar and Baddeley 1984; Vallar and Cappa 1987). The former is a passive input phonological store, the latter an active articulatory rehearsal process. Phonologically coded information can be held in the 116 Journal of Neurolim, Volume 7, Number 112(1992) phonological store, and refreshed by the articulatory rehearsal in order to prevent trace decay; on the other hand, the articulatory rehearsal feeds phonologically recoded visual information to the store. According to the working memory model, articulatory processes play a crucial role in the short-term retention of verbal material, and this assumption is consistent with many previous experimental studies (Levy 1971; Murray 1975). Recently, some clinical and neuropsychological studies on patients affected by motor speech disturbances have explored possible consequences of selective artlculatory impairment (Nebes 1975; Levine et zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH al. 1982; Villa and Caltagirone 1984; Baddeley and Wilson 1985; Vallar and Cappa 1987; Rochon er al. in press). All patients were totally mute or had severe speech-out disturbances, due to cortical, subcortical, or brain-stem lesions. Almost all studies reported normal memory span even in patients who could not produce overt speech. Only Rochon et al. (in press) found a consistently lower verbal span in five left-hemisphere damaged patients affected by so-called “speech apraxia”. Taken together, these studies suggested that “peripheric” impairments of articulation did not disrupt articulatory rehearsal, and were not sufficient to determine a verbal memory deficit. By contrast, “central” disorders of articulation, as well as those in apraxic patients, were thought to provoke short-term deficits, though not necessarily excluding articulatory rehearsal. The papers cited above usually collected together patients who had different anatomical lesions (both “peripheric”, i.e. pontine lesions and “central”, cortical damage), and whose data at additional tasks (phonological similarity, and word length effect) were not consistent with each other. Our research was aimed at verification of this theoretical claim in a more homogeneous group of patients with non-linguistic motor speech disturbances. Motor speech disturbances are generally divided into two subgroups (Darley er al. 1975): (i) apraxia of speech, a disorder of the “capacity to program” speech muscular movements, due to cortical damage; (ii) dysarthria, a group of different speech disorders with failure of muscular “control” and “strength”, not exclusively affecting speech production. On the basis of speech-out features, it is possible to distinguish six types of dysarthria, differing in lesion site and pathogenesis (Dailey et al. 1969a, b): flaccid, spastic, hypokinetic, hyperkinetic, ataxic, and mixed. The above papers on verbal short-term memory described patients affected by apraxia of speech or by spastic, flaccid, or mixed dysarthria. Instead, we chose patients affected by ataxic dysarthria, an articulatory derangement due to altered sequences of incorrect movements. Such a picture is a symptom of cerebellar dysfunction, independent of cortical lesions, and of any specific linguistic deficit. In particular, we studied patients affected by hereditary ataxias, a group of rare diseases, with selective degeneration of spino-cerebellar system and cerebellum, characterized by ataxia (i.e. failure in coordination) of the extremities, nystagmus, and dysarthria. Heredoataxic patients represent a sufficiently homogeneous group of patients suffering from impaired coordination and control of articulation, phonation, and respiration. oUrstudyconsistedoffourexperimentpaimedtoascertainwhetherttrcse”pcriferic”, selective, well characterized, homogeneous, motor speech disturbances could affect the articulatory loop. SUBJECTS We examined a continuous series of twenty-five out-patients affected by heredodegenerative ataxias. Eighteen subjects were selected without neuropsychological (Token Test; and Raven’s Coloured Matrices), and neuroimaging (CT scan) evidence of cerebral cortiud damage. Inclusion criteria were: more than 28/36 at Token Test; more than 22/36 at Raven’s Coloured Matrices. Seven out of 25 patients (28%) were excluded, and this figure can be thought of an an indirect expression of the prevalence of dementia in heredoataxias. Some authors consider it “rare” (like Slaby and Wyatt 1974), while other authors reported very much higher figures (e.g. Werdelin 1986; but there dementia was roughly defined). Thirteen patients (seven male, six female) received a diagnosis of Freidreich’s disease according to the Ametee Cooperative Study on Friedreich’s disease (Geoftroy et al, 1976). The mean age was 22.6 years (range 14-33, S.D. 6.4), mean education was 9.2 years (range 5-13, S.D. 2.2). Three patients (two males, one female) received a diagnosis of Autosomal Dominant Cerebellar Ataxia (A.D.C.A.), type 1, according to Harding (1983). Mean age was 46 years (range 35-55), mean education was 9.1 years (range 8-12). Two patients (two males) had Early Onset Cerebellar Ataxia with retained tendon reflexes (E.O.C.A.), as defined by Harding (1983). Mean age was 28.5 years (range 24-32), mean education was 8 years in both of them. Only four of these patients (age range 16-30; schooling range 8-13), three affectad by Friedreich’s disease (two male, one female) and one by E.O.C.A., accepted to perform Experiment 2, which was given as a supplemental task. Severity of the disease was evaluated by Inherited Ataxias Progression Scale (I.A.P.S.; Campanella et zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJI al. 1980), which varies from stage I @symptomatic patient) to stage IV (patient confined to wheelchair or to bed). Our Friedreich’s disease patients were grouped into stages II (six subjects), III (three subjects), and IV (four subjects); A.D.C.A. and E.O.C.A. patients were included in stage II. Mean duration of illness was 10.7 years (range 2-22) in Friedreich patients, 22 years (range 8-18) in A.D.C.A., and 26 years (range 22-30) in E.O.C.A. subjects. Dysarthria is found to be present in early stages of both Friedreich’s disease and hereditary cerebellar ataxias, and in many cases it can be a debut symptom (Weredelm 1986). All of our patients showed dysarthria to a different degree: from difficulties 118 Journal of Neurdingulstics, Volume 7, Number 112(1992) and hesitations in pronouncing some complex phonemes to a picture of overt dysarmria. Fifty normal volunteers, without neurologic or psychiatric diseases, were examined as controls. Age (tti = 0.1; p = 0.9) and schooling (t = 0.9; p = 0.30) matched those of patients considered as a whole group (Table 1). TABLE 1 Subjects N Patients Controls 18 50 Schooling (years) Age (Y-W M S.D. M S.D. 27.1 27.5 11.0 10.0 9.1 9.9 2.8 3.2 Patients’ and controls’ mean age and education, with standard deviation (S.D.). Thirteen (seven male, six female) patients were affected by Friedreich’s disease; three patients (two male, one female) by autosomal dominant cerebellar ataxia; two patients (male) by early onset cere-bellar ataxia. EXPERIMENT 1: IMMEDIATE AND DELAYED WITHOUT ARTICULATORY SUPPRESSION RECALL WITH AND In the exploration of articulatory loop in dysarthric patients, one should first evaluate verbal immediate memory abilities, and verify if patients use the same strategy as control subjects. The present experiment had this two-fold aim: it investigated patients’ immediate recall and the effect of articulatory suppression on delayed recall. Method The experiment consisted of three different conditions of serial recall. In each condition, twenty sequences of three bisyllabic high-frequency concrete words were presented auditorily at a standard rate of one item/s. (a) In the first condition, subjects were asked to repeat items immediately after auditory presentation. Subjects had to reproduce sequences in correct serial order, and in the case of omissions, they were invited to specify the serial position of items recalled. (b) In the second condition, serial recall was requested after a 6-s free interval. (c) In the third, the 6-s interval was filled with articulatory suppression (i.e. subjects had to count forward starting from 1 at a regular, quick rate). In these two tasks, recall instructions were the same as in (a). The score was the total number of items recalled in correct serial position for each condition (maximum score 60). All 18 patients were given this test, and Experiments 3and4. R& zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA a nd Disc ussion Results are given in Table 2. Patients scored lower in all three conditions, but they performed similarly to controls. Patients’ and controls’ best performance was at the first condition: both groups showed a decrease in delayed recall (second condition), and a greater drop with articulatory suppression. Two-way ANOVA showed evidence of significant group (& ,& = 38.46,~ = 0.0001) and task (Qj2 = 198.69; p = O.ooOl) effects. Posr-hoc comparisons were performed by means of the Scheffe’ test. Within group analysis was consistent with the idea of parallel performance by patients and controls. In both groups, no significant difference was found between the first and second condition (patients: zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJ F = 0.15, and controls: F = 1.59; p = N.S.), while the differences between the first and third (patients: F = 68.21, and controls: F = 91.05; p < 0.001). and between the second and third condition (patients: F = 61.84, and controls: F = 68.54; p < 0.001) were very significant. TABLE 2 Immediate zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA and Delayed Verbal Recall with and witbout articalatory suppreadon First Cond. Patients Controls Second Cond. Third Cond. M S.D. M S.D. M S.D. 56.9 59.4 2.9 1.6 55.6 57.3 3.7 3.0 30.2 43.9 11.0 9.4 Input ordered recall in three conditions: immediate (First Con&), a!& a 6-s free delay (Second Cond.); after a 6-s delay filled with articulation of irrelevant material (articulatory suppression: Third Cond.). Means (M) and standard deviations (S.D.) arc shown; maximum score: 60. Comparisons between groups showed a significant difference between patients and controls at the first condition (F = 21.04, p = 0.001). suggesting that patients had a deficit in verbal immediate memory (mean difference between patients* and controls’ scores: 2.51). Interestingly, only borderline significance was found for patientcontrol comparison at the second condition (F = 3.86; p = 0.54), where patients showed a less evident decrease (mean difference: 1.72). Finally, in the third condition, patient-control difference was clearly significant (F = 25.71; p = 0.001); in particular, under articulatory suppression, patients showed a very marked drop in recall 120 Journal of Neurolingu&tics,Volume 7, Number 112(1992) accuracy (mean difference: 13.67). Second condition results could account for significant group-task interaction (F4,,ss = 590.0; p = 0.0001). The present experiment indicated a verbal memory deficit in dysarthric patients, but suggested that patients did rehearse items, as well as normal controls. Actually, if dysarthric patients had chosen a somewhat different strategy, avoiding rehearsal because of their articulatory troubles, articulatoty suppression would have been ineffective, with similar results at the second and third condition. Vice versa, articulatory suppression produced a marked detrimental effect, while no difference was found between immediate recall and the free-interval condition. In other words, patients could not be considered as normal subjects under articulatory suppression, nohvithstanding their articulatory troubles. These findings were highly suggestive, but did not definitely prove that patients were using articulatory rehearsal. Two observations were worthwhile. The former was concerned with the unexpected reduction of the mean difference between patients and controls at the second condition. Both groups showed an insignificant decrement in performance at delayed recall, but controls dropped more evidently. We are not able to explain the controls’ results, but delayed recall findings supported the view that both groups were using the same strategy. The latter consideration regarded the very marked patients’ breakdown under articulatory suppression. At least two interpretations are possible. The former would foresee that patients suffered a greater disadvantage from performing two tasks at same time (memory and articulation tasks), independently of their nature, just as patients with “Central Executive System” deficit (Baddeley et al. 1986). A second interpretation would foresee that the detrimental effect of articulatory suppression was specifically dependent on the articulatory nature of the task. In this case, irrelevant articulation could be more expensive for patients in terms of attentional resource demand, because of the speech output difficulties. In other terms, patients’ lower verbal span could be linked to reduced availability of attentional resources during articulation. However, this interpretation would not follow directly theoretical claims of working memory, according to which articulatory rehearsal has minimal attentional demand. Alternatively, patients’ breakdown under suppression could be due to the complete impossibility to rehearse, while normal subjects may “intersperse an occasional subvocal rehearsal between suppression responses” (Baddeley 1986). Additional, indepedent data are necessary to address this issue properly. A last remark about the present and following experiments: in contrast to previous studies in dysarthric patients, we did not test serial recall by writing or pointing because all patients were able to give oral responses. Also the possible implications of this choice will be discussed below. ArtkulatoryLoopinAtukDgsutbrh zyxwvutsrqponmlkjihg 121 EXPERIMENT 2: ARTICULATORY LOOP EFFECTS To confirm that patients used articulatory rehearsal, we looked for effects generally considered to be due to an articulatoty component of the immediate verbal memory system: the phonological similarity effect for visually presented material and word length effect for both auditory and visual stimuli (Vallar and Cappa 1987). A pattern clearly indicative of disruption of articulator-y rehearsal was found in only one (B.O.) out of the five apraxic patients with verbal span deficit as described by Rochon et al. (in press). Of the remaining four patients, two showed standard effects of visual phonological similarity and word length (indicative of normal articulatory rehearsal), while two (A. M. and G. B.) did not show clear patterns. Also Vallar and Cappa’s paper (1987) reported contrasting results in two anarthric patients, suggesting a spared articulatory rehearsal with normal verbal span. By contrast, and strikingly, the pattern expected in the case of articulatory rehearsal disruption was found, and attributed to a peculiar strategy choice, in a patient (P. V.) with defective PS and reduced verbal span (Vallar and Baddeley 1984). Method We searched for phonological similarity effect and for word length effect by means of the same procedure already used in Italian studies (Vallar and Baddeley 1984; Vallar and Cappa 1987). Only four out of eighteen patients accepted to perform this experiment, as described in a previous section. (a) Phonological similarity effect: we used two pools of letters (10 phonologically similar and 10 phonologically dissimilar) to build up sequences of three and four items. Ten different sequences from each set were both aurally and visually presented, and patients had to recall items immediately afterwards. The score was the total number of items repeated in correct serial position (maximum score thirty for three item sequences and forty for four item sequences). (b) Word length effect: we used two sets of words (10 bisyllabic and 10 pentasyllabic) to build up sequences of three and four items. Ten different sequences from each set were both aurally and visually presented, and patients had to repeat them immediately afterwards. The score was the total number of items repeated in correct serial position (maximum score was thirty for three item sequences, and forty for four item sequences). RESULTS AND DISCUSSION Results, reported in Tables 3 and 4, are expressed as percentages of maximum score. About phonological similarity effect, a higher accuracy in reproducing dis- 122 Journal of Neudh@da, Volume 7, Number 112(1992) similar phonemes was observed both with shorter and longer sequences, independently of presentation modality. A two-way ANOVA showed that phonological similarity approached significance (F, ,, = 4.1; p = 0.052), while modality (F,,, = 1.8; p = 0.18) and similarity-modality interaction (F,,*s = 0.35; p = 0.56) did not. In this case, results were clear-cut even though a fully significant phonological similarity effect was not found. According to the theoretical framework, articulatory rehearsal conveys phonologically recoded visual information to the phonological store, thus determining phonological similarity effect for visual material. In this case, the insignificant modality effect and, in particular, the insignificant similaritymodality interaction testified that phonological similarity effect occurred to the same extent with both auditory and visual presentation, i.e. dysarthric patients were using articulatory rehearsal. TABLE 3 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR phonological similarity effect Visual pres . Acoustic pres . Length Ph.S. Ph.D. Ph.S. Ph.D. 3 4 0.95 0.71 0.99 0.92 0.89 0.67 0.97 0.81 Input ordered recall of sequences of three or four phonologically similar (Ph.S.) and phonologically dissimilar letters, presented aurally (acoustic pres.) and visually (visual pres.). Results are expressed as mean probability of correct recall. Four patients were given this task. TABLE 4 Word Length Effect Visual pres. Acoustic pres . 3 4 Short Long Short Long 0.99 0.85 0.95 0.73 0.95 0.85 0.91 0.66 Input ordered recall of sequences of three or four short and long words, aurally (acoustic pres.) and visually (visual pres.) presented. Results are expressed as mean probability of correct recall. Four patients were given this task. As for the word length effect, a similar analysis was done. A higher overall accuracy for short words was evident, both with three item and four item sequences, independently of presentation modality. A two-way ANOVA showed a significant word length AcWWoryLoopinAtuicDy- 123 effect (F,., = 4.5; zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA p = O&l), while there was not a modality effect (Ft.7 = 0.9; 0.35), nor a word length-modality interaction (F,,,, = 0.803; p = O.%). That is to say a clear word length detrimental effect was found in verbal immediate memory. In addition, the lack of modality effect and of length-modality interaction was consistent with the view that our patients were not under articulatory suppression. Otherwise, suppression being present during input, it would have been possible to observe differences between auditory and visual presentation (Baddeley et al. 1984). Once more, our findings suggested that patients were using articulatory rehearsal, and in this case evidence seemed even stronger. In synthesis, dysarthric patients behaved differently from normal subjects under articulatory suppression. They showed the typical effects attributed both to articulator-y rehearsal and to phonological store, as well as normal Italian speakers (Vallar and Baddeley 1984), and in general as well as normal subjects. Pure motor speech disorders did not disrupt articulatory loop. It must be noted that this conclusion is valid both for patients considered as a whole group, and for single patients. Actually, all subjects presented clear phonological similarity and word length effects, when patients did not achieve ceiling results. p = EXPERIMENT 3: ARTICULATION RATE The articulatory loop was found to be unimpaired. Another feature of speech output relevant to the issue of immediate memory is articulation rate, and the present experiment was aimed at measuring it. In previous studies it was implicitly assumed that articulatory rehearsal features overlapped those of overt realization of speech. For instance, normal articulation rate testified for normal rehearsal processes in the patient P. V., who showed a selective impairment of phonological store (Baddeley and Vallar 1984). Moreover, systematic relationships between articulation time and verbal memory abilities were found: the faster the articulation rate, the larger the verbal memory span (Baddeley et al. 1975; Nicolson 1981; Schweickert and Boruff 1986; Standing and Curtis 1989). Method We used two simple measures of articulation flow, already described in Italian speakers (Baddeley and Vallar 1984). We measured time spent in: (a) counting forwards from 1 to 10 five successive times; and (b) articulating letters of the Italian alphabet from A to Z. Subjects were asked to articulate aloud as quickly as possible, pronouncing items correctly. 124 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Journal of Neurolinguistics,Volume 7, Number 112(19912) Results zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA and Discussion As expected, patients’ mean articulation time was very high (Table S), even though the lowest patients’ results overlapped into a normal range. The next question is whether these findings are sufficient to account for a defective immediate serial recall. In other words, can patients’ verbal memory deficit be thought of as the clinical counterpart of articulation time-verbal span relationships? This claim would be in agreement with the model and also with the idea that articulatory rehearsal is involved in memory span (Hitch 1980; Vallar and Papagno 1986). However, even though we have selected patients without neuropsychological deficits, it remains plausible that the verbal memory deficit is part of a more general deficit of short-term memory, seen either as a defect of the “Central Executive” or as a consequence of a decreased information processing speed (Hart et al. 1986). TABLE 5 Articulation Rate Alphabet Patients Controls Numbers M S.D. Range M SD. Range 23.2 11.0 13.6 2.6 10-75 8-18 10.2 5.3 9.1 1.5 8-35 3-10 Means (M), standard deviations (S.D.), and range of articulation times (in s) for Italian Alphabet (A-Z) and numbers (l-10; five times successively). Since patients have been shown to rehearse at a slower rate, it could be possible that the testing procedure, which implied oral responses, could account for patients’ lower results in immediate verbal recall. Actually, a study on prefix/sufftx effect (Baddeley and Hull 1979) pointed out that the longer the time after stimulus the presentation is, the poorer the retention, and oral responses induced a larger delay between presentation and response completion in patients, because they took a longer time to pronounce items. Indeed, contrasting data have been reported on the effect of the delay between presentation and response (for a discussion, see Frick 1988). Therefore, we suggest that larger time delays after presentation can be considered as a minor factor in verbal memory performance. EXPERIMENT 4: IMMEDIATE MEMORY SPANS To ascertain whether our patients showed a general short-term memory defect or a selective verbal one, we gave patients verbal and nonverbal memory span tests. We chose a nonverbal test, the Corsi’s block tapping test, which, according to the working memory model, is believed to tap a different slave system, devoted to visuo-spatial immediate memory. We tested memory for both digits and words. (a) The Verbal Memory Span for digits was given and scored according to the standard instructions of the W.A.I.S. Digits Forward subtest (Wechsler 1981), which is well standardixed in Italian speakers (Orsini et al. 1987). (b) The Verbal Memory Span for words was included to assess verbal span with different material. We asked subjects to repeat series of bisyllabic high-fquency concrete words. Three series for each sequence length were given. The score was the longest sequence the patient could repeat correctly. Such a further measure of verbal span has already been used in Italian clinical studies (Orsini et al. 1988). (c) Furthermore, we used Corsi’s block tapping task, following the standard procedure (Orsini et (II. 1987). Results are shown in Table 6. Patients’ mean score was lower both at the verbal and spatial span, but score ranges overlapped. One-way ANOVA applied to data for each independent variable (digit span, word span, Corsi’s span) showed that only digit span (F,,, = 4.40; p = 0.04) and word span (F,,& = 7.49; zyxwvutsrqponmlkjihgfedcbaZYXWV p = 0.008) dif f er ed significantly in the two groups, while spatial span difference failed to reach significance (F,,, = 3.5; p = 0.09). Corsi Patients COnaOlS Digits words M S.D. M S.D. lu S.D. 4.8 5.2 0.8 0.9 4.9 5.4 0.8 0.9 3.9 4.5 0.5 0.9 Means (hi) and standad deviations (S.D.) for immediatemmrory span. The present experiment was aimed at ascertaining whether these patients had a selective verbal short-term memory, or a more general defti of short-term memory. NsL 7:1/2-a 126 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Journalof NewoIinguistics,Volume 7, Number 112(1992) From the theoretical point of view, a selective verbal STM defect would rule out the hypothesis of a Central Executive defect in these patients, formulated to explain the relevant drop of delayed recall under articulatory suppression. Our span results are not strong enough to allow definite conclusions. Nonetheless, since spatial span, which relies more heavily on “Central Executive”, both in encoding and decoding processes (Morris 1987), appears to be less affected in our patients, it is possible to suggest that the defect of “Central Executive” at least is less likely. Actually, an opposite pattern (selective deficit at Corsi’s block tapping task with spared verbal immediate memory) was found in patients with defective “Central Executive”, such as mildly demented patients (Orsini er al. 1988). Consequently, even though we cannot exclude a slight, general short-term memory defect, we think that the more prominent verbal memory defect has to be related to a more specific mechanism. The slower articulation rate in dysarthric patients could induce a slight verbal memory defect per se. To verity this interpretation we searched for correlations between articulatory rate and short-term memory tasks. As a measure of articulation rate, we used the mean of the two articulation times (articulation index) for each subject. We also tested the eventual correlation between the articulation index and spatial span, to exclude the possibility that articulation index could relate to STM defect non-specifically. Correlation coefficients are shown in Table 7 for patients and controls separately. TABLE 7 Correlation Coef’ficients Between Articulatory Index and Memory Span Patients Controls Digit span word span zyxwvutsrqponmlkjihgfedcbaZYX Spatial span r P r P r P 0.57 0.3 0.01 0.03 0.28 0.06 0.25 0.65 0.16 0.14 0.5 0.33 All controls’ coefficients were lower, probably because of the lower variability in their articulation index. Articulatory index was not correlated with spatial span, while it was significantly correlated with digit span in both groups. However, the correlation of articulatory index with word span only approached significance for patients, while it was not significant for controls. So, on one hand it seems clear that articulation time is not only an index of general cognitive impairment, while on the other hand, it seems to relate well to digit span while it was not highly correlated with word serial recall task. This last finding, i.e. the lack of correlation between articulatory index and word span, could be explained by the heterogeneity of the material to be pronounced (digits, letters) and to be recalled (words). Indeed, studies about articulation-span Articulatory Loop in Ataxic Dysvthrip 127 relationships used the same verbal material for articulating and recalling. In summary, memory span results could not provide firm evidence supporting a “pure” verbal immediate memory defect, but it is possible to suggest that slower articulatory rate plays a valuable role in patients’ verbal span deficit. GENERAL DISCUSSION The aim of this paper was to verify possible influences of motor speech disturbances upon articulatory loop. We selected patients presenting ataxic dysarthria without general cognitive impairments and investigated their performance. The outstanding result was that articulatory loop was unimpaired in this homogeneous group of subjects. Patients did use the same strategy in verbal immediite recall as normal controls, and showed all standard effects of phonological similarity, and of wordlength for both auditory and visual presentation. Nevertheless, dysarthric patients showed slow articulation rate and a slight deficit of immediate memory, more prominent for verbal material. It was argued, also on the basis of a significant negative correlation between articulation times and verbal span, that the low articulation rate can at least contribute to, if not account for, the verbal memory deficit. This hypothesis has the merit of being entirely compatible with all previous experimental studies and it is consistent with normal articulatory loop effects, found in ataxic patients. Actually, one could hypothesize that patients were still using articulatory rehearsal but with reduced efficiency depending on its rate. In other words, if articulatory loop is able to maintain verbal material that can be pronounced in approximately 2 s (Baddeley et al. 1975), dysarthric patients can be seen to have only a “capacity limitation” of articulatory rehearsal. Analogously, patient P.V. was shown to have only a “capacity limitation” of a functional phonological store (Baddeley and Vallar 1984), because she presented the standard phonological similarity effect, and normal articulation rate, but had a reduced auditory verbal short-term memory. In both cases, auditory phonological similarity and word length effects represent qualitative indices of preserved articulatory rehearsal and phonological store, respectively, whereas they cannot detect capacity limitations or reduced efficiency of the subcomponents. Consequently, it would not be possible to rule out an impairment of phonological store in our patients only on the basis of standard detrimental effect of auditory phonological similarity. In other terms, subjects could have a phonological store capacity limitation, not detectable with articulatory loop effects, that would be sufficient to explain the reduction of verbal memory span. However, at a glance it would seem unlikely that patients with subcortical lesions could show a phonological store capacity limitation. Furthermore, our dysarthric patients differed from patients with defective phonological store by a constant, though not significant, superiority of 128 Journal of Neurolinguistics, Volume 7, Number l/2 (1992) auditory modality in all verbal memory tasks and for normal comprehension abilities (as assessed by means of the Token Test). On the basis of our results, we confirm that articulatory rehearsal operates by means of “central” components, because articulatory troubles due to non-cortical disorders did not affect articulatory rehearsal processes. However, we found that our patients did show slight verbal memory deficit, in contrast to patients affected by more severe speech-out disorders previously described. These inconsistencies deserve some consideration. Patients with massive pontine lesions are reported to show normal verbal span, but only on a pointing task, with standard effects of word length and phonological similarity. As discussed above, standard effects tell us nothing about the features, accuracy, and efficiency of the “inner voice”. That is to say, standard articulatory loop effects cannot make quantitative predictions. The verbal memory impairment we found, has been ascribed to motor speech disturbances due to sub-tentorial selective lesions of the coordination and control (cerebellar) pathways. In other patients, defined as “peripheric”, as G. F. (Vallar and Cappa 1987), massive pontine lesions involved efferent pathways but probably also control pathways (cerebellar) and general activating cortical afferences (reticular midbrain formation). Therefore, these patients ’ “inner voices” could present alterations of accuracy and rate sufficient per se to provoke deficits in verbal immediate memory. This deficit would be impossible to detect, but standard articulatory loop effects are not a sufficient argument for excluding its existence. From this point of view, we suggest some cautions about claims of normal verbal STM in patients with gross disturbances of overt articulation. Only more detailed assessment of each subject, or, on the other hand, group studies of homogeneous patients can allow for such general conclusions about unimpaired short-term memory. In conclusion, “peripheric” motor speech disorders of ataxic dysarthria do not affect articulatory loop subcomponents. Nonetheless, it was argued that the slower articulation rate due to articulatory troubles can represent a kind of “capacity limitation” of articulatory rehearsal and can account for (at least partially) a slight verbal immediate memory deficit. From a methodological point of view, the notion of partial impairments (“capacity limitations”) of single subcomponents, as well as that of possible exceptions to the model conceivable in terms of strategical choices (likewise in patients G. F., and P. V.), could extend the theoretical framework and make it more compliant. However, the more flexible and comprehensive a framework becomes, the less falsifiable it becomes, and hence it is only with strong, convergent evidence that theoretical exceptions are allowed into a framework. Bearing in mind these considerations, we were very cautious about drawing conclusions which are entirely compatible with previous experimental data and with theoretical constraints. Otherwise we had to admit that the data did not fit into working memory theory. Artkuhtory Loop in Ataxk zyxwvutsrqponmlkjihgfedcbaZY Dysduia zyxwvutsrqponm 129 zyxwvu NOTES 1. Address for correspondence: Dr Luigi Trojano, Fondazione Clinica de1 Lavoro, Centro Medico di Campoli M.T., I.R.C.C.S., 82030 Campoli M.T. (BN) Italy. REFERENCES Baddeley, A. D. 1986 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA W orking M emory , Oxford: Clarendon Press. Baddeley, A. D. and G. J. Hitch 1974 “Working Memory,” in Recent Advances in Learning and M otivation, Vol. VIII, Bower G. (ed.), New York: Academic Press. Baddeley, A. D. and A. Hull 1979 “Prefix and Suffix Effect: Do They Have a Common Basis?” Journal of Verbal Learning and Verbal Behavior 18. 12940. Baddeley, A. D., V. Lewis and G. Vallar 1984 “Exploring the Articulatory Loop,” Quarterly Journal of Experimental Psy chology 36. 233- 52. Baddeley, A. D., R. Logie, S. Bressi, S. Della Sala and H. Spinnler 1986 “Dementia and Working Memory,” Quarterly Journal of Experimental Psy chology 38a. 603- 18. Baddeley, A. D., N. Thomson and M. Buchanan 1975 “Word Length and the Structure of Short-term Memory,‘* Journal of Verbal Learning and Verbal Behavior 14. 575- 89. Baddeley, A. D. and B. Wilson 1985 “Phonological Coding and Short-term Memory in Patients Without Speech,” Journal of M emory and Language 24. 490- 502. Campanella, G., A. Filla, F. De Falco, D. Mansi, A. Durivage and A. Barbeau 1980 “Friedreich’s Ataxia in the South of Italy. A Clinical and Biochemical Survey of 23 Patients,” Canadian Journal of Neurological Sciences 7. 351- 8. Darley, F. L., A. E. Aaronson and J. R. Brown 1969a “Differential Diagnostic Patterns of Dysarthria,” Journal of Speech and Hearing Research 12, 146- 69. 1969b “Clusters of Deviant Speech Dimentions in Dysarthrias,” Journal of Speech and Hearing Research 12. 462- 9. 1975 M otor Speech Disorders, Toronto: W. B. Saunders. Frick, R. W. 1988 “Issue of Representation and Limited Capacity in the Auditory Short-term Store,” British Journal of Psy chology 79. 2 1340. 130 Journal of Neurolinguistics, Volume 7, Number 112 (1992) Geoffroy, G., A. Barbeau, G. Breton, B. Lenieux, M. Aube, C. Leger and J. B. Bouchard 1976 “Clinical Description and Roentgenologic Evaluation of Patients with Friedreich’s Ataxia,” zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON Canadian Journal of Neurological Sciences 3. 279- 86. Harding, A. E. 1. 1983 “Classification of the Hereditary Ataxias and Paraplegias,” Lancer zyxwvutsrq 1151-5. Hart, R. P., G. K. Henry, J. A. Kwentus and R. T. Leshner 1986 “Information Processing Speed of Children with Friedreich’s Ataxia,” Developmental M edicine and Child Neurology 20. 310- 3. Hitch, G. J. 1980 “Developing the Concept of Working Memory,” in Cognitive Psy chology : New Directions, Claxton G. (ed.), London: Routledge &LKegan Paul. Levine, D. N., R. Calvanio and A. Popovics 1982 “Language in the Absence of Inner Speech,” Neuropsy chologia 20. 391- 409. B. A. 1971 “Role of Articulation in Auditory and Visual Short-term Memory,” Journal of Verbal Learning and Verbal Behavior 10. 123-32. Morris, R. G. 1987 “Exploring the Visuo-Spatial Scratch-pad,*’ Quarterly Journal of Experi- Levy , mental Psy chology 39. 409- 30. Murray, D. J. 1975 “Articulation and Acoustical Confusability in Short-term Memory,” Journal of Experimental Psy chology 78. 679- 84. Nebes, R. D. 1975 “The Nature of Internal Speech in a Patient with Aphemia,” Brain and Language 2. 489- 97. Nicolson. R. 1981 “The Relationship Between Memory Span and Processing Speed,” in Intelligence and Learning, M. Friedman, J. P. Das and N. O’Connor (eds), London: Plenum Press. Orsini. A., D. Grossi, E. Capitani, M. Laiacona, C. Papagno and G. Vallar 1987 “Verbal and Spatial Immediate Memory Span: Normative Data from 1355 Adults and I I12 Children,” Italian Journal of Neurological Sciences 8. 539- 48. Orsini, A., L. Trojano, L. Chiacchio and D. Grossi 1988 “Immediate Memory Spans in Dementia,” Perceptual and M oror Skills 67. 267- 72. Al- tkUhtO ~Lo o piDAtUkDyssrthrlr zyxwvutsrqponmlkji 131 Rochon, E., G. Waters and D. Caplan n.d. “Short-term Memory Processes in Patients with Apraxia of Speech: Implications for the Nature and Structure of the Auditory Verbal Shortterm Memory System,” Journal of Neuroiinguistic in press. Schweickert, R. and B. Boruff 1986 “Short-term Memory Capacity: Magic Number or Magic Spell?” Journal of Experimental Psychology: Learning, Memory, and Cognition 12. 419-25. Slaby, A. E. and R. J. Wyatt 1974 Dementia in the Presenium, Springfield (Jll.): C. C. Thomas Publishers. Standing, L. and L. Curtis 1989 “Subvocalization Rate Versus Other Predictors of the Memory Span,” Psychological Repot% 65. 487-95. Vallar, G. and A. D. Baddeley 1984 “Fractionation of Working Memory: Neuropsychological Evidence for a Phonological Short-term Store,” Journal of Verbal Learning and Verbal Behavior 23. 151-61. Vallar. G. and S. F. Cappa 1987 “Articulation and Verbal Short-term Memory: Evidence from Anarthria,” Cognitive Neuropsychoiogy 4. 55-78. Vallar. G. and C. Papagno 1986 “Phonological Short-term Store and the Nature of the Recency Effect: Evidence from Neuropsychology,” Brain and Cognition 5. 428-42. Villa, G. and C. Caltagirone 1984 “Speech Suppression Without Aphasia after Bilateral Perisylvian Softenings,” Italian Journal of Neurological Sciences 5. 77-83. Wechsler, D. 1981 WA&R Manual, New York: Psychological Corporation. Werdelin, L. 1986 “Hereditary Ataxias. Occurrence and Clinical Features,*’ Acta Neurologica Scandinavica 73. Supp. 106.