Marialuisa Martelli

Are faces processed like words? A diagnostic test for recognition by parts

Do we identify an object as a whole or by its parts? This simple question has been surprisingly hard to answer. It has been suggested that faces are recognized as wholes and words are recognized by parts. Here we answer the question by applying a test for crowding. In crowding, a target is harder to identify in the presence of nearby flankers. Previous work has described crowding between objects. We show that crowding also occurs between the parts of an object. Such internal crowding severely impairs perception, identification, and fMRI face-area activation. We apply a diagnostic test for crowding to a word and a face, and we find that the critical spacing of the parts required for recognition is proportional to distance from fixation and independent of size and kind. The critical spacing defines an isolation field around the target. Some objects can be recognized only when each part is isolated from the rest of the object by the critical spacing. In that case, recognition is by parts. Recognition is holistic if the observer can recognize the object even when the whole object fits within a critical spacing. Such an object has only one part. Multiple parts within an isolation field will crowd each other and spoil recognition. To assess the robustness of the crowding test, we manipulated familiarity through inversion and the face- and word-superiority effects. We find that threshold contrast for word and face identification is the product of two factors: familiarity and crowding. Familiarity increases sensitivity by a factor of x1.5, independent of eccentricity, while crowding attenuates sensitivity more and more as eccentricity increases. Our findings show that observers process words and faces in much the same way: The effects of familiarity and crowding do not distinguish between them. Words and faces are both recognized by parts, and their parts -- letters and facial features -- are recognized holistically. We propose that internal crowding be taken as the signature of recognition by parts.

Length Effect in Word Naming in Reading: Role of Reading Experience and Reading Deficit in Italian Readers

Vocal reaction times (RTs) in naming 3- to 8-letter words were measured in proficient and dyslexic readers (Study 1). In proficient readers, RTs were independent of word length up to 5-letter words, indicating parallel processing. In the 5- to 8-letter range, RTs increased linearly, indicating sequential processing. Reading experience was associated with both faster discrimination of individual elements and parallel processing of increasingly large word parts. In dyslexics, RTs increased linearly with increasing length indicating reliance on sequential decoding. Individual analysis indicated 2 profiles of RTs (Types A and B). In Study 2, the distinction between A and B dyslexics was not associated with the use of different reading procedures. However, a more marked speed deficit characterized Type B dyslexics.

Flicker flutter: Is an illusory event as good as the real thing?

Verghese and Stone (1995) showed that reducing the perceived number of objects by grouping also reduces objective performance. Shams, Kamitani, and Shimojo (2000) showed that a single flash accompanied by multiple beeps appears to flash more than once. We show that objective orientation-discrimination performance depends solely on the perceived number of flashes, independent of the actual number of beeps and flashes. Thus the unit of perceptual analysis seems to be a perceived event, independent of how it is induced.

Bridging the gap between different measures of the reading speed deficit in developmental dyslexia

The study assessed how decoding and pronunciation times contribute to total reading time in reading aloud and how these measures change in the presence of developmental dyslexia. Vocal reaction times (RTs), pronunciation times, and total reading times were measured while 25 children with dyslexia and 43 age-matched typically developing readers read singly presented words and non-words that varied for length. Group differences were large for vocal RTs; children with dyslexia were increasingly slower as a function of condition difficulty (over-additivity effect); lexicality and length influenced RTs even when over-additivity was controlled for by z-score transformation. The group differences were also large for vocal total reading times, but the effect of over-additivity was smaller than that of vocal RTs and no selective influence of lexicality and length was detected. Pronunciation times showed very small individual differences and no over-additivity effect; children with dyslexia were more sensitive to the effect of lexicality and length than controls. To assess the contribution of the cognitive and sensory–motor compartments in determining group differences, we applied the difference engine model. As for RTs, the relationship between means and standard deviations closely supported the prediction of a general cognitive delay in the slow group, with no group difference in the sensory–motor compartment. The variance in total reading times was predicted by combining the model results for RTs with the linear relationship between pronunciation times and task difficulty. The results help clarify the internal structure of reading times, a measure largely used in clinical testing to assess reading rate.

Two different mechanisms for omission and substitution errors in neglect dyslexia

Neglect dyslexia is a reading disorder often associated with right-sided brain lesions. In reading single words, errors are mostly substitutions or omissions of letters that occupy the left-sided positions. Typically, these errors have been thought to depend on a single mechanism. Conversely, we propose that they are due to different mechanisms. In particular, a visuo-spatial mechanism is responsible for omissions and a perceptual integration process for substitution errors. We measured the performance of six patients with both neglect and neglect dyslexia, analyzing their reading errors as a function of letter spacing. According to our conjecture, letter spacing should increase omissions by moving part of the string further in the unattended space, while it should reduce substitutions by restoring the integration processes. Furthermore, we predict that letter spacing should be more effective with pseudowords compared to words, in that in this latter case lexical effects are supposed to influence attentional and perceptual processes. Accordingly, we found that for pseudowords only the two types of errors are differently affected by this manipulation and only omissions correlate with the severity of the disorder in visuo-spatial tasks.

Neglect dyslexia: a matter of "good looking".

Brain-damaged patients with right-sided unilateral spatial neglect (USN) often make left-sided errors in reading single words or pseudowords (neglect dyslexia, ND). We propose that both left neglect and low fixation accuracy account for reading errors in neglect dyslexia. Eye movements were recorded in USN patients with (ND+) and without (ND-) neglect dyslexia and in a matched control group of right brain-damaged patients without neglect (USN-). Unlike ND- and controls, ND+ patients showed left lateralized omission errors and a distorted eye movement pattern in both a reading aloud task and a non-verbal saccadic task. During reading, the total number of fixations was larger in these patients independent of visual hemispace, and most fixations were inaccurate. Similarly, in the saccadic task only ND+ patients were unable to reach the moving dot. A third experiment addressed the nature of the left lateralization in reading error distribution by simulating neglect dyslexia in ND- patients. ND- and USN- patients had to perform a speeded reading-at-threshold task that did not allow for eye movements. When stimulus exploration was prevented, ND- patients, but not controls, produced a pattern of errors similar to that of ND+ with unlimited exposure time (e.g., left-sided errors). We conclude that neglect dyslexia reading errors may arise in USN patients as a consequence of an additional and independent deficit unrelated to the orthographic material. In particular, the presence of an altered oculo-motor pattern, preventing the automatic execution of the fine saccadic eye movements involved in reading, uncovers, in USN patients, the attentional bias also in reading single centrally presented words.

Dissociation in Optokinetic Stimulation Sensitivity between Omission and Substitution Reading Errors in Neglect Dyslexia

Although omission and substitution errors in neglect dyslexia (ND) patients have always been considered as different manifestations of the same acquired reading disorder, recently, we proposed a new dual mechanism model. While omissions are related to the exploratory disorder which characterizes unilateral spatial neglect (USN), substitutions are due to a perceptual integration mechanism. A consequence of this hypothesis is that specific training for omission-type ND patients would aim at restoring the oculo-motor scanning and should not improve reading in substitution-type ND. With this aim we administered an optokinetic stimulation (OKS) to two brain-damaged patients with both USN and ND, MA and EP, who showed ND mainly characterized by omissions and substitutions, respectively. MA also showed an impairment in oculo-motor behavior with a non-reading task, while EP did not. The two patients presented a dissociation with respect to their sensitivity to OKS, so that, as expected, MA was positively affected, while EP was not. Our results confirm a dissociation between the two mechanisms underlying omission and substitution reading errors in ND patients. Moreover, they suggest that such a dissociation could possibly be extended to the effectiveness of rehabilitative procedures, and that patients who mainly omit contralesional-sided letters would benefit from OKS.

Perceptual and Cognitive Factors Imposing “Speed Limits” on Reading Rate: A Study with the Rapid Serial Visual Presentation

Adults read at high speed, but estimates of their reading rate vary greatly, i.e., from 100 to 1500 words per minute (wpm). This discrepancy is likely due to different recording methods and to the different perceptual and cognitive processes involved in specific test conditions. The present study investigated the origins of these notable differences in RSVP reading rate (RR). In six experiments we investigated the role of many different perceptual and cognitive variables. The presence of a mask caused a steep decline in reading rate, with an estimated masking cost of about 200 wpm. When the decoding process was isolated, RR approached values of 1200 wpm. When the number of stimuli exceeded the short-term memory span, RR decreased to 800 wpm. The semantic context contributed to reading speed only by a factor of 1.4. Finally, eye movements imposed an upper limit on RR (around 300 wpm). Overall, data indicate a speed limit of 300 wpm, which corresponds to the time needed for eye movement execution, i.e., the most time consuming mechanism. Results reconcile differences in reading rates reported by different laboratories and thus provide suggestions for targeting different components of reading rate.

Channel for reading

Letter identification is mediated by just one spatial frequency channel (Solomon and Pelli, 1994). But what about reading? We wondered whether larger features, e.g. words, at lower spatial frequencies are used when reading text. The demo shows text on a background of narrowband noise varying in spatial frequency from top to bottom. Try reading it. You read quickly at the top, slow down in the middle, and speed up again at the bottom. The middle frequencies impair reading most, just as found with letter identification (Majaj et al. 2002). No channel tuned to words was revealed.

Subtypes of developmental dyslexia in transparent orthographies: a comment on Lachmann and Van Leeuwen (2008).

Lachmann and Van Leeuwen (2008) proposed two diagnostic subtypes of developmental dyslexia in a language with transparent orthography (German). The classification was based on reading time, rather than reading errors, for lists of words and nonwords. The two subtypes were “frequent-word reading impaired” (FWRI) and “nonword reading impaired” (NWRI). Notably, FWRI were very slow in reading high-frequency words but as fast as controls in reading nonwords; ca. one-third of these children showed this “reversed lexicality effect” in a particularly marked fashion (i.e., read nonwords two to three times faster than high-frequency words). Since Italian is a highly transparent language, we applied this classification to 87 third- and sixth-grade dyslexics from various previously published studies. Some children showed a marked lexicality effect, while others showed small or no difference between word and nonword reading speed. However, regardless of stimulus length, grade and presence/absence of a previous language delay, no child showed a marked reversed lexicality effect; more generally, no child could be classified as FWRI. These findings indicate that the search for subtypes of developmental dyslexia in transparent orthographies still constitutes an open question.