Made by DATEXIS (Data Science and Text-based Information Systems) at Beuth University of Applied Sciences Berlin
Deep Learning Technology: Sebastian Arnold, Betty van Aken, Paul Grundmann, Felix A. Gers and Alexander Löser. Learning Contextualized Document Representations for Healthcare Answer Retrieval. The Web Conference 2020 (WWW'20)
Funded by The Federal Ministry for Economic Affairs and Energy; Grant: 01MD19013D, Smart-MD Project, Digital Technologies
"Aphasia is usually first recognized by the physician who treats the person for his or her brain injury. Most individuals will undergo a magnetic resonance imaging (MRI) or computed tomography (CT) scan to confirm the presence of a brain injury and to identify its precise location." In circumstances where a person is showing possible signs of aphasia, the physician will refer him or her to a speech-language pathologist (SLP) for a comprehensive speech and language evaluation. SLPs will examine the individual's ability to express him or herself through speech, understand language in written and spoken forms, write independently, and perform socially.
The American Speech, Language, Hearing Association (ASHA) states a comprehensive assessment should be conducted in order to analyze the patient's communication functioning on multiple levels; as well as the effect of possible communication deficits on activities of daily living. Typical components of an aphasia assessment include: case history, self report, oral-motor examination, language skills, identification of environmental and personal factors, and the assessment results. A comprehensive aphasia assessment includes both formal and informal measures.
Formal assessments:
- Boston Diagnostic Aphasia Examination (BDAE): diagnoses the presence and type of aphasia, focusing on location of lesion and the underlying linguistic processes.
- Western Aphasia Battery - Revised (WAB): determines the presence, severity, and type of aphasia; and can also determine baseline abilities of patient.
- Communication Activities of Daily Living - Second Edition (CADL-2): measures functional communication abilities; focuses on reading, writing, social interactions, and varying levels of communication.
- Revised Token Test (RTT): assess receptive language and auditory comprehension; focuses on patient's ability to follow directions.
Informal Assessments:
Informal assessments aide in the diagnosis of patients with suspected aphasia.
- Conversational Speech and Language Sample
- Family Interview
- Case History or Medical Chart Review
- Behavioral Observations
Diagnostic information should be scored and analyzed appropriately. Treatment plans and individual goals should be developed based on diagnostic information, as well as patient and caregiver needs, desires, and priorities.
The best way to see if anomic aphasia has developed is by using verbal as well as imaging tests. The combination of the two tests seem to be most effective, since either test done alone may give false positives or false negatives. For example, the verbal test is used to see if there is a speech disorder, and whether it is a problem in speech production or in comprehension. Patients with Alzheimer's disease have speech problems that are linked to dementia or progressive aphasias which can include anomia. The imaging test, mostly done using MRI scans, is ideal for lesion mapping or viewing deterioration in the brain. However, imaging cannot diagnose anomia on its own because the lesions may not be located deep enough to damage the white matter or damage the arcuate fasciculus. However, anomic aphasia is very difficult to associate with a specific lesion location in the brain. Therefore, the combination of speech tests and imaging tests has the highest sensitivity and specificity.
It is important to first do a hearing test, in case the patient cannot clearly hear the words or sentences needed in the speech repetition test. In the speech tests, the person is asked to repeat a sentence with common words; if the person cannot identify the word but he or she can describe it, then the person is highly likely to have anomic aphasia. However, to be completely sure, the test is given while a test subject is in an fMRI scanner, and the exact location of the lesions and areas activated by speech are pinpointed. Few simpler or cheaper options are available, so lesion mapping and speech repetition tests are the main ways of diagnosing anomic aphasia.
Sensory aphasia cannot be diagnosed through the use of imaging techniques. Differences in cognition between asymptomatic subjects and affected patients can be observed via functional magnetic resonance imaging (fMRI). However, these results only reveal temporal differences in cognition between control and diagnosed subjects. The degree of progression during therapy can also be surveyed through cognition tests monitored by fMRI. Many patients’ progress is assessed over time via repeated testing and corresponding cerebral imaging by fMRI.
There are few neuropsychological assessments that can definitively diagnose prosopagnosia. One commonly used test is the famous faces tests, where individuals are asked to recognize the faces of famous persons. However, this test is difficult to standardize. The Benton Facial Recognition Test (BFRT) is another test used by neuropsychologists to assess face recognition skills. Individuals are presented with a target face above six test faces and are asked to identify which test face matches the target face. The images are cropped to eliminate hair and clothes, as many people with prosopagnosia use hair and clothing cues to recognize faces. Both male and female faces are used during the test. For the first six items only one test face matches the target face; during the next seven items, three of the test faces match the target faces and the poses are different. The reliability of the BFRT was questioned when a study conducted by Duchaine and Nakayama showed that the average score for 11 self-reported prosopagnosics was within the normal range.
The test may be useful for identifying patients with apperceptive prosopagnosia, since this is mainly a matching test and they are unable to recognize both familiar and unfamiliar faces. They would be unable to pass the test. It would not be useful in diagnosing patients with associative prosopagnosia since they are able to match faces.
The Cambridge Face Memory Test (CFMT) was developed by Duchaine and Nakayama to better diagnose people with prosopagnosia. This test initially presents individuals with three images each of six different target faces. They are then presented with many three-image series, which contain one image of a target face and two distracters. Duchaine and Nakayama showed that the CFMT is more accurate and efficient than previous tests in diagnosing patients with prosopagnosia. Their study compared the two tests and 75% of patients were diagnosed by the CFMT, while only 25% of patients were diagnosed by the BFRT. However, similar to the BFRT, patients are being asked to essentially match unfamiliar faces, as they are seen only briefly at the start of the test. The test is not currently widely used and will need further testing before it can be considered reliable.
The 20-item Prosopagnosia Index (PI20) is a freely available and validated self-report questionnaire that is able to identify individuals with prosopagnosia. It has been validated against the famous faces test and Cambridge Face Memory Test, with evidence that PI20 scores are correlated with performance on these objective measures of face recognition. It can be downloaded from the Royal Society's Open Science website and on . Alternatively, the questionnaire can be completed online on the official website.
Sensory aphasia is typically diagnosed by non-invasive evaluations. Neurologists, neuropsychologists or speech pathologists will administer oral evaluations to determine the extent of a patient’s comprehension and speech capability. Initial assessment will determine if the cause of linguistic deficiency is aphasia. If the diagnosis is then confirmed, testing will next address the type of aphasia and its severity. The Boston Diagnostic Aphasia Examination specializes in determining the severity of a sensory aphasia through the observation of conversational behaviors. Several modalities of perception and response are observed in conjunction with the subject’s ability to process sensory information. The location of the brain lesion and type of the aphasia can then be inferred from the observed symptoms. The Minnesota Test for Differential Diagnosis is the most lengthy and thorough assessment of sensory aphasia. It pinpoints weaknesses in the auditory and visual senses, as well as reading comprehension. From this differential diagnosis, a patient’s course of treatment can be determined. After treatment planning, the Porch Index of Communicative Ability is used to evaluate prognosis and the degree of recovery.
In order to assess an individual for agnosia, it must be verified that the individual is not suffering from a loss of sensation, and that both their language abilities and intelligence are intact. In order for an individual to be diagnosed with agnosia, they must only be experiencing a sensory deficit in a single modality. To make a diagnosis, the distinction between apperceptive and associative agnosia must be made. This distinction can be made by having the individual complete copying and matching tasks. If the individual is suffering from a form of apperceptive agnosia they will not be able to match two stimuli that are identical in appearance. In contrast, if an individual is suffering from a form of associative agnosia, they will not be able to match different examples of a stimulus. For example, an individual who has been diagnosed with associative agnosia in the visual modality would not be able to match pictures of a laptop that is open with a laptop that is closed.
Using verbal descriptions may be helpful for individuals with certain types of agnosia. Individuals such as prosopagnosics may find it useful to listen to a description of their friend or family member and recognize them based on this description more easily than through visual cues.
There are tests that can indicate with high probability whether a person is a dyslexic. If diagnostic testing indicates that a person may be dyslexic, such tests are often followed up with a full diagnostic assessment to determine the extent and nature of the disorder. Tests can be administered by a teacher or computer. Some test results indicate how to carry out teaching strategies.
Someone with jargon aphasia may exhibit the following behaviors[4]:
- Intermixing real words and nonsensical words while speaking or writing
- Failing to recognize mistakes being made while speaking or writing
- Using real words in incorrect situations
- Frequent, repetitive uttering of low frequency words
- The inability to say or write a specific word or phrase
Some of the specific types of language errors that occur are[5]:
Lexical (real word):
Semantic - Real word that was semantically related to target. ("dog" instead of "cat")
Formal - Real word that shared either the initial phoneme or at least 50% of phonemes with target. ("dog" instead of "desk" or "dog" instead of "frog")
Mixed - Real word that was both semantically and phonologically related to target. ("bicycle" instead of "motorcycle")
Visual - Real word of an item similar in visual form to the target. ("ball" instead of "orange")
Unrelated - Real word that was not related to the target in any obvious way. ("dog" instead of "apple")
Non-lexical (nonword):
Phonological - Nonword that shared either the initial phoneme or at least 50% of phonemes with target. ("deg" instead of "dog")
Neologistic - Nonword not reaching the criterion for phonological relatedness (i.e., sharing less than 50% of phonemes with the target and with a different initial phoneme). Nonwords that are pseudo compound words. ("kib" instead of "dog")
Other Errors:
Don’t know - Indication that response was unknown or if item was not responded to at all. ("I don’t know" or silence)
Description - Attempts to describe as opposed to name item. (Multiple word responses)
Jargon aphasia must be diagnosed through a series of tests. Since the number of individuals that have aphasia after suffering a stroke is high, a test is usually carried out soon after the stroke occurs. There is a list of basic exercises to help assess a person’s language skills, such as:
- Naming objects that begin with a certain letter
- Reading or writing
- Holding a conversation
- Understanding directions and commands
There is also a common test used, called the Boston Diagnostic Aphasia Examination test, which incorporates exercises that extensively review the person’s language skills[2].
Management strategies for acquired prosopagnosia, such as a person who has difficulty recognizing people's faces after a stroke, generally have a low rate of success. Acquired prosopagnosia sometimes spontaneously resolves on its own.
Agraphia cannot be directly treated, but individuals can be rehabilitated to regain some of their previous writing abilities.
For the management of phonological agraphia, individuals are trained to memorize key words, such as a familiar name or object, that can then help them form the grapheme for that phoneme. Management of allographic agraphia can be as simple as having alphabet cards so the individual can write legibly by copying the correct letter shapes. There are few rehabilitation methods for apraxic agraphia; if the individual has considerably better hand control and movement with typing than they do with handwriting, then they can use technological devices. Texting and typing do not require the same technical movements that handwriting does; for these technological methods, only spatial location of the fingers to type is required. If copying skills are preserved in an individual with apraxic agraphia, repeated copying may help shift from the highly intentional and monitored hand movements indicative of apraxic agraphia to a more automated control.
Micrographia is a condition that can occur with the development of other disorders, such as Parkinson's disease, and is when handwriting becomes illegible because of small writing. For some individuals, a simple command to write bigger eliminates the issue.
- Anagram and Copy Treatment (ACT) uses the arrangement of component letters of target words and then repeated copying of the target word. This is similar to the CART; the main difference is that the target words for ACT are specific to the individual. Target words that are important in the life of the individual are emphasized because people with deep or global agraphias do not typically have the same memory for the words as other people with agraphia may. Writing can be even more important to these people as it can cue spoken language. ACT helps in this by facilitating the relearning of a set of personally relevant written words for use in communication.
- Copy and Recall Treatment (CART) method helps to reestablish the ability to spell specific words that are learned through repeated copying and recall of target words. CART is more likely to be successful in treating lexical agraphia when a few words are trained to mastery than when a large group of unrelated words is trained. Words chosen can be individualized to the patient, which makes treatment more personalized.
- Graphemic buffer uses the training of specific words to improve spelling. Cueing hierarchies and copy and recall method of specific words are used, to work the words into the short-term memory loop, or graphemic buffer. The segmentation of longer words into shorter syllables helps bring words into short-term memory.
- Problem solving approach is used as a self-correcting method for phonological errors. The individual sounds out the word and attempts to spell it, typically using an electronic dictionary-type device that indicates correct spelling. This method takes advantage of the preserved sound-to-letter correspondences when they are intact. This approach may improve access to spelling memory, strengthen orthographic representations, or both.
Dyslexic children require special instruction for word analysis and spelling from an early age. While there are fonts that may help people with dyslexia better understand writing, this might simply be due to the added spacing between words. The prognosis, generally speaking, is positive for individuals who are identified in childhood and receive support from friends and family.
Specialists, like ophthalmologists or audiologists, can test for perceptual abilities. Detailed testing is conducted, using specially formulated assessment materials, and referrals to neurological specialists is recommended to support a diagnosis via brain imaging or recording techniques. The separate stages of information processing in the object recognition model are often used to localize the processing level of the deficit.
Testing usually consists of object identification and perception tasks including:
- object-naming tasks
- object categorization or figure matching
- drawing or copying real objects or images or illustrations
- unusual views tests
- overlapping line drawings
- partially degraded or fragmented image identification
- face or feature analysis
- fine line judgment
- figure contour tracking
- visual object description
- object-function miming
- tactile ability tests (naming by touch)
- auditory presentation identification
Sensory modality testing allows practitioners to assess for generalized versus specific deficits, distinguishing visual agnosias from optic aphasia, which is a more generalized deficit in semantic knowledge for objects that spans multiple sensory modalities, indicating an impairment in the semantic representations themselves.
At its most basic level, dyscalculia is a learning disability affecting the normal development of arithmetic skills.
A consensus has not yet been reached on appropriate diagnostic criteria for dyscalculia. Mathematics is a specific domain that is complex (i.e. includes many different processes, such as arithmetic, algebra, word problems, geometry, etc.) and cumulative (i.e. the processes build on each other such that mastery of an advanced skill requires mastery of many basic skills). Thus dyscalculia can be diagnosed using different criteria, and frequently is; this variety in diagnostic criteria leads to variability in identified samples, and thus variability in research findings regarding dyscalculia.
Other than using achievement tests as diagnostic criteria, researchers often rely on domain-specific tests (i.e. tests of working memory, executive function, inhibition, intelligence, etc.) and teacher evaluations to create a more comprehensive diagnosis. Alternatively, fMRI research has shown that the brains of the neurotypical children can be reliably distinguished from the brains of the dyscalculic children based on the activation in the prefrontal cortex. However, due to the cost and time limitations associated with brain and neural research, these methods will likely not be incorporated into diagnostic criteria despite their effectiveness.
According to Bates et al. (2005), "the primary goal of rehabilitation is to prevent complications, minimize impairments, and maximize function". The topics of intensity and timing of intervention are widely debated across various fields. Results are contradictory: some studies indicate better outcomes with early intervention, while other studies indicate starting therapy too early may be detrimental to the patient's recovery. Recent research suggests, that therapy be functional and focus on communication goals that are appropriate for the patient's individual lifestyle.
Specific treatment considerations for working with individuals with Wernicke's aphasia (or those who exhibit deficits in auditory comprehension) include using familiar materials,using shorter and slower utterances when speaking, giving direct instructions, and using repetition as needed.
Neuroplasticity: Role in Recovery
Neuroplasticity is defined as the brain's ability to reorganize itself, lay new pathways, and rearrange existing ones, as a result of experience. Neuronal changes after damage to the brain such as collateral sprouting, increased activation of the homologous areas, and map extension demonstrate the brain's neuroplastic abilities. According to Thomson, "Portions of the right hemisphere, extended left brain sites, or both have been shown to be recruited to perform language functions after brain damage. All of the neuronal changes recruit areas not originally or directly responsible for large portions of linguistic processing. Principles of neuroplasticity have been proven effective in neurorehabilitation after damage to the brain. These principles include: incorporating multiple modalities into treatment to create stronger neural connections, using stimuli that evoke positive emotion, linking concepts with simultaneous and related presentations, and finding the appropriate intensity and duration of treatment for each individual patient.
There is no method available to completely cure anomic aphasia. However, there are treatments to help improve word-finding skills.
Although a person with anomia may find it difficult to recall many types of words such as common nouns, proper nouns, verbs, etc., many studies have shown that treatment for object words, or nouns, has shown promise in rehabilitation research. The treatment includes visual aids, such as pictures, and the patient is asked to identify the object or activity. However, if that is not possible, then the patient is shown the same picture surrounded by words associated with the object or activity. Throughout the process, positive encouragement is provided. The treatment shows an increase in word-finding during treatment; however, word identifying decreased two weeks after the rehabilitation period. Therefore, it shows that rehabilitation effort needs to be continuous for word-finding abilities to improve from the baseline. The studies show that verbs are harder to recall or repeat, even with rehabilitation.
Other methods in treating anomic aphasia include Circumlocution Induced Naming therapy (CIN), wherein the patient uses circumlocution to assist with his or her naming rather than just being told to name the item pictured after given some sort of cue. Results suggest that the patient does better in properly naming objects when undergoing this therapy because CIN strengthens the weakened link between semantics and phonology for patients with anomia, since they often know what an object is used for but cannot verbally name it.
Anomia is often burdensome on the families and friends of those suffering from it. One way to overcome this burden is computer-based treatment models, effective especially when used with clinical therapy. Leemann et al. provided anomic patients with computerized-assisted therapy (CAT) sessions, along with traditional therapy sessions using treatment lists of words. Some of the patients received a drug known to help relieve symptoms of anomia (levodopa) while others received a placebo. The researchers found that the drug had no significant effects on improvement with the treatment lists, but almost all of the patients improved after the CAT sessions. They concluded that this form of computerized treatment is effective in increasing naming abilities in anomic patients.
Additionally, one study researched the effects of using "excitatory (anodal) transcranial direct current stimulation" over the right temporo-parietal cortex, a brain area that seems to correlate to language. The electrical stimulation seemed to enhance language training outcome in patients with chronic aphasia.
1. SCAN is the most common tool for diagnosing APD, and it also standardized. It is composed for four subsets: discrimination of monaurally presented single words against background noise, acoustically degraded single words, dichotically presented single words, sentence stimuli. Different versions of the test are used depending on the age of the patient.
2. Random Gap Detection Test (RGDT) is also a standardized test. It assesses an individual’s gap detection threshold of tones and white noise. The exam includes stimuli at four different frequencies (500, 1000, 2000, and 4000 Hz) and white noise clicks of 50 ms duration. It is a useful test because it provides an index of auditory temporal resolution. In children, an overall gap detection threshold greater than 20 ms means they have failed.
3. Gaps in Noise Test (GIN) also measures temporal resolution by testing the patient's gap detection threshold in white noise.
4. Pitch Patterns Sequence Test (PPT) and Duration Patterns Sequence Test (DPT) measure auditory pattern identification. The PPS has s series of three tones presented at either of two pitches (high or low). Meanwhile, the DPS has a series of three tones that vary in duration rather than pitch (long or short). Patients are then asked to describe the pattern of pitches presented.
The only way to treat aphasia is with speech and language therapy (SLT). It will not completely restore the person’s prior level of communication, but SLT can lead to a massive improvement of jargon aphasia. Recipients of this treatment typically achieve better use of residual language abilities, improved language skills, and the ability to communicate in a different way by making up for missing words in their speech[2].
One specific method that has shown to lead to improvements with certain symptoms is “phonological component analysis”, or PCA for short. Participants in PCA therapy tend to improve in the ability to name specific items that they are test on, as well as the decrease in use of nonwords to describe said items. Seeing promising results from this type of therapy has led to much optimism in hopes of developing more treatment methods for jargon aphasia[6].
Though there have been ample attempts to rehabilitate patients with pure alexia, few have proven to be effective on a large scale. Most rehabilitation practices have been specialized to a single patient or small patient group. At the simplest level, patients seeking rehabilitation are asked to practice reading words aloud repeatedly. This is meant to stimulate the damaged system of the brain. This is known as multiple oral re-reading (MOR) treatment. This is a text-based approach that is implemented in order to prevent patients from LBL reading. MOR works by reading aloud the same text repeatedly until certain criteria are reached. The most important criteria for a pure alexic patient is reading at an improved rate. The treatment aims to shift patients away from the LBL reading strategy by strengthening links between visual input and the associated orthographic representations. This repetition supports the idea of using top-down processing initially minimize the effects peripheral processing which were demonstrated in the study above. From here, the goal is to increasing bottom-up processing. This will hopefully aid in word recognition and promote interactive processing of all available information to support reading. 'The supported reading stimulation from MOR has a rehabilitative effect so that reading rate and accuracy are better for untrained text, and word-form recognition improves as evidenced by a reduced word-length effect.' These tactics have seen quite good success.
Another tactic that has been employed is the use of cross modal therapy. In this therapy, patients are asked to trace the words in which they are trying to read aloud. There has been success using cross modal therapy such as kinaesthetic or motor-cross cuing therapy, but tends to be a more feasible approach for those on the slower reading end of the spectrum.
The nature of the alleged mental representations that underlie the act of pointing to target body parts have been a controversial issue. Originally, it was diagnosed as the effects of general mental deterioration or of aphasia on the task of pointing to body parts on verbal command. However, contemporary neuropsychological therapy seeks to establish the independence of autotopagnosia from other disorders. With such a general definition, a patient that presents with a dysfunction of or failure in accessing one of four mental representation systems suffers from autotopagnosia. Through observational testing, the type of mental misrepresentation of the body can be deduced: whether "semantic", "visuospatial", "somatosensory", or "motor misrepresentations". Neuropsychological tests can provide a proper diagnosis in regards to the specificity of patient’s agnosic condition.
1) Test 1: Body Part Localization: Free vision and no vision conditions
2) Test 2: On-line positioning of body vis-à-vis objects
3) Test 3: Localization of objects on the body surface
4) Test 4: Body part semantic knowledge
5) Test 5: Matching body parts: Effect of viewing angle
Transient paraphasias (as well as other language defects such as speech arrest) can be generated by artificially activating the brain's language network with Transcranial magnetic stimulation (TMS). With navigated TMS (nTMS), nodes of the language network can be located presurgically so that critical areas can be saved when performing tumor or epilepsy surgery. Marketed by Nexstim, this method has received Food and Drug Administration (FDA) clearance in the United States.
Phonological dyslexia is a reading disability that is a form of alexia (acquired dyslexia), resulting from brain injury, stroke, or progressive illness and that affects previously acquired reading abilities. The major distinguishing symptom of acquired phonological dyslexia is that a selective impairment of the ability to read pronounceable non-words occurs although the ability to read familiar words is not affected. It has also been found that the ability to read non-words can be improved if the non-words belong to a family of pseudohomophones.
Many language impairments, including paraphasic errors, are reduced in number through spontaneous recovery of neurological function; this occurs most often with stroke patients within the first three months of recovery. Lesions associated with ischemic strokes have a shorter spontaneous recovery time, within the first two weeks, and lesions associated with hemorrhagic strokes, on the other hand have a longer period for spontaneous recovery, four to eight weeks. Whether spontaneous recovery occurs or not, treatment must begin immediately after the stroke. A traditional approach requires treatment beginning at the level of breakdown - in the case of paraphasia, at the level of the phoneme. There are commercially available workbooks that provide various activities such as letter, word-picture, or word-word matching, and sentence completion, among other things. The difficulty of these activities varies with the level of treatment. However, these treatments have not been proven to be clinically productive. Functional magnetic resonance imaging is the most widely used technique to study treatment-induced recovery, looking at activation of particular areas of the brain. There are many different ways to process fMRI scans, beginning with the pre-scanning process. Data must be normalized. There is also no consensus on whether or not single subject scans are more helpful than group scans to determine a general pattern of treatment. However, fMRI scans have a few disadvantages.
A 1988 study by Mary Boyle proposed a method focused on oral reading to treat phonemic paraphasias was partially successful, resulting in fewer phonemic paraphasias but a slower rate of speech. Treatments lasted for 50 minutes and occurred once a week. During these treatment sessions, the patient was instructed to look at twenty different phrases -each of these phrases consisted of one to three syllables - then read the phrase. If the patient failed to read the phrase, the process was repeated. If the patient failed to read the phrase again, the process was abandoned. To progress from a set of one syllable phrases to two syllable phrases and two syllable phrases to three syllable phrases, an 80% success rate was necessary. This treatment was partially successful. Although fewer phonemic paraphasias were produced due to this treatment, speaking efficiency was not improved by this study. This is partially because the focus of the treatment was on sound production rather than semantic content. Improvements lasted for six weeks before the patient regressed.
Surface dyslexia is a type of dyslexia, or reading disorder. According to Marshall & Newcombe's (1973) and McCarthy & Warrington's study (1990), patients with this kind of disorder cannot recognize a word as a whole due to the damage of the left parietal or temporal lobe. Individuals with surface dyslexia are unable to recognize a word as a whole word and retrieve its pronunciation from memory. Rather, individuals with surface dyslexia rely on pronunciation rules. Thus, patients with this particular type of reading disorder read non-words fluently, like "yatchet", but struggle with words that defy pronunciation rules (i.e. exception words). For example, a patient with surface dyslexia can correctly read regular words like "mint", but will err when presented a word that disobeys typical pronunciation rules, like "pint". Often, semantic knowledge is preserved in individuals with surface dyslexia.