Friday, July 29, 2016

Autism in Adults: A Survey - Part 6: Functional Language

This is part 6 of a series on a survey I did in 2012. To get background information, go to part 1. To read an analysis of executive dysfunction in the sample, go to part 2. To read an analysis of alexithymia in autism, go to part 3. To read an analysis of the results from a new questionnaire assessing reaction to eye contact, go to part 4, and to read an analysis of a new questionnaire about autism internal experience, go to part 5. Lastly, to read about independent living skills, go to part 7.

In this section, I discuss responses to several qualitative questions discussing communication ability.

First, I asked about baseline language ability. Written language appeared to be better than speech in both receptive and expressive domains, with 20.5% reporting difficulty with receptive speech and 29.5% with expressive speech, compared to only 9.3% reporting difficulty with reading and 2.3% with writing. Overall, 61.9% had normal baseline language ability, 23.8% had impairment in one area and 14.3% had impairment in two areas, with no participants having three or more areas of impairment. There were significant correlations between expressive and receptive spoken language (r = .413, p = .005) and receptive spoken and written ability (r = .464, p = .002).

Age was correlated with receptive speech (r = .471, p = .002), with an ANOVA finding that 17-24 year olds were more likely to have impairments in receptive language than 36-60 year olds (25-35 year olds were intermediate and didn't significantly differ from either age group). Specifically, 47.1% of 17-24 year olds, 16.7% of 25-35 year olds and none of the 36-60 year olds reported baseline receptive speech issues. In addition, female participants were more likely to have receptive speech issues, with 26.1% of women and 6.2% of men having baseline receptive speech impairment.

In addition, 25% of individuals reported significant variability in their expressive language skills, while 22.7% reported variability in receptive language. In most cases, the trigger for this variability seemed to be overload, emotional distress, fatigue, or distracting stimuli (such as background noise or eye contact). The variability could range from a subjective feeling of greater communication difficulty in certain situations, to becoming temporarily nonverbal. Variability in written language was less common, with only two people reporting variability in writing ability and none in reading ability.

In this sample, 11.4% had variability in expressive or receptive language ability and 18.2% had variability in both receptive and expressive language ability (usually with the same conditions impairing both). Of the two participants with variable writing ability, both lost written and spoken expressive ability under the same conditions. Variability in receptive and expressive language was significantly correlated (r = .689, p = .001). Variability in receptive and expressive language were both more common in women, with 34.8% of women having variable expressive language and 30.4% having variable receptive compared to 6.3% and 12.5% of men.

The individuals with variability in language performance were not necessarily those with poor baseline performance, with no correlation between baseline performance and variability found. Considering all six language problems described so far, only 45.2% of my sample had consistently normal language ability, with 21.4% having one language issue, 26.2% having two language issues and 7.1% having three language issues. None of my sample had more than three of these issues.

There were no correlations between language impairments and AQ total or Reaction to Eye Contact factor scores. However, the only individual with impaired baseline written language had the lowest score on AQ imagination of the sample (scoring 2), and individuals with impaired baseline receptive spoken language scored lower on AQ communication (r = .339, p = .030).

Poor baseline receptive speech was also associated with higher FrSBe total scores (r = .322, p = .043). Variable receptive language was associated with higher scores on FrSBe total (r = .357, p = .030), disinhibition (r = .351, p = .027) and executive dysfunction (r = .327, p = .042). This suggests that individuals with greater frontal lobe-related problems also tended to have poorer receptive language and a tendency to lose receptive language under stress.

Lastly, baseline receptive speech was correlated with Autism Internal Experience factor 1, with individuals with impaired receptive speech scoring higher (r = .311, p = .045). The impaired group scored an average of 17.75+/-3.105, with the unimpaired group scoring 13.09+/-6.147. High scorers on this factor were visuospatial thinkers who tended not to use internal language. This suggests a link between receptive language and verbal thinking, which has been suggested by previous studies as well. This study found that autistic individuals who scored better on nonverbal IQ than verbal IQ were less likely to use inner speech while solving math problems than autistics with other profiles.

It's interesting to speculate about the reasons for better and more consistent written than spoken language performance in my participants. This may reflect the written modality of the survey, meaning that individuals with impaired written language were less likely to complete my survey. Alternately, it could reflect better visual than auditory language abilities (particularly since AIE factor 1, reflecting visual thinking and lack of internal speech, was correlated with receptive spoken language impairment). It could also be due to greater processing time allowed by written communication. It would have been interesting to see if signing is easier than speech, but unfortunately extensive childhood exposure to sign language is not typical for most autistic adults. (Hearing autistic children of Deaf adults would be able to shed light on that question, but they're a minority within a minority.)

However, many studies into autistic children with complex communication needs have found that sign language can be useful for many children - this study successfully taught a few signs to autistic kids who had made little progress learning speech, this study found that more children acquired receptive signing than receptive speech from exposure to SimCom (simultaneous communication - speaking and signing the same words at the same time), and this study found that three minimally verbal autistic children readily learnt signs from key word signing but only two of them showed improvements in speech. Although these studies all used participants with much more severe language problems than my sample, this suggests that the better written than spoken performance in my sample may be due to the visual modality of communication.

In addition, receptive language problems appear to be more common in younger participants. This may suggest that these issues lessen with age, that certain ages have reporting biases, or that younger generations of autistic individuals are more likely to have receptive speech issues. Very few studies have assessed changes in language ability between adolescence and middle age in autism. The only study I could find reported no significant change in verbal IQ in a longitudinal sample assessed in adolescence and in adulthood, although both adolescent and adult scores were higher than childhood scores. Please note that it's unclear whether the mild language issues described by my participants would translate into lowered VIQ, especially since VIQ combines multiple areas of language into one score. In particular, in adolescents and adults, VIQ assesses both written and spoken language.

Women also had more severe receptive language problems than men, both in terms of baseline performance and variability. This runs counter to the general tendency for girls to be better at language than boys, however, other studies have also found autistic females tend to have more language problems than autistic males. Why this should be the case is unclear.

Variability in receptive language appears linked to frontal lobe issues. Anecdotally, sensory overload or disruptive sensory cues (such as background noise) were a common trigger for this variability - something that other studies have found as well.. The ability to screen out irrelevant stimuli is sometimes called 'sensory gating', and has been linked to the prefrontal cortex in both individuals with schizophrenia and people with brain injuries. In particular, the second study found that auditory distractions disproportionately affected the performance of patients with prefrontal brain lesions (as opposed to lesions in other regions) on an auditory matching task.



Although these language issues are mild compared to the issues experienced by some autistic individuals (such as individuals with complex communication needs), they were all significant enough to be noticed by the individual, and likely all caused some level of problems with communication on a regular basis. There is a stereotype that Asperger Syndrome causes no language problems, and high functioning autism causes language problems only in early childhood. However, even individuals who are not currently showing a noticeable language delay can have more subtle difficulties in language - difficulties which may be easily overlooked by others.

Further research should be done. This data suggests that many autistic adults may have better written than spoken language skills, subtle receptive and/or expressive language problems, and/or substantial situational variability in language functioning. These difficulties may be linked to frontal lobe issues, visual thinking and sensory processing issues.

Monday, July 18, 2016

Autism in Adults: A Survey - Part 5: Internal Experience

This is part 5 of a series on a survey I did in 2012. To get background information, go to part 1. To read an analysis of executive dysfunction in the sample, go to part 2. To read an analysis of alexithymia in autism, go to part 3. To read an analysis of the results from a new questionnaire assessing reaction to eye contact, go to part 4. To read about functional language, go to part 6, and lastly, to read about independent living skills, go to part 7.

In section, I discuss the second of the two new measures I used in this study - a measure asking about a variety of commonly anecdotally described internal experiences in autism, on a 40-item, three-point scale.

First, I decided to see which items correlated with AQ total. Surprisingly, only items 3, 4, 27, 32 and 34 significantly correlated with AQ total. Item 4 was negatively correlated, with an r of -.439 and a p of = .014. The other four items were positively correlated, with r ranging from .381-.602 and p from .035 to <.001. These items described individuals who had difficulty believing others had different likes and dislikes, had trouble knowing their own feelings, found words related to their interests made them happy, found it difficult to visualize faces and found their own face didn't look as they expected.

I then analyzed which items correlated with each AQ subscale. AQ social was correlated with items 4, 16, 27 and 32. Item 4 was negatively correlated with AQ social (r = -.455, p = .004) and the other three items were all positively correlated (r =.363-.511, p = .023-.001). These items described individuals who had difficulty knowing their own feelings, visualized a sentence that NTs don't typically visualize, found words related to their interest made them happy, and had difficulty visualizing faces.

AQ attention-switching was correlated with items 3, 10, 27 and 29. All items were positively correlated, with r ranging from .342-.407 and p from .029-.008. These items described individuals who had difficulty believing others could have different likes and dislikes from their own, heard a person's voice in their head while remembering what they said, found words related to their interest made them happy and visualized numbers as occurring in a line going from smaller to larger.

AQ attention to detail was correlated with items 8, 18, 20, 24, 27 and 34. All items were positively correlated, with r ranging from .331-.532 and p from .040 to <.001. These items described individuals who visualized written text while being spoken to, had letter-color synesthesia, found that certain places cued certain thoughts, felt like events in different locations belonged to different timelines, found words related to their interest made them happy and found their own face didn't look as they expected.

AQ communication was correlated with items 1, 2, 3, 4, 15, 26, 30 and 34. Items 4 and 26 were negatively correlated, with r ranging from -.426 to -.489 and p from .006-.001. The other items were positively correlated, with r ranging from .332-.537 and p ranging from .036 to <.001. These items described individuals who had to consciously translate thoughts to words, represented concepts nonverbally and couldn't translate some thoughts into words at all. In addition, these individuals had difficulty knowing how they were feeling, difficulty believing others could have different likes and dislikes, tended not to notice hunger while concentrating, and found their own face didn't look as they expected.

AQ imagination was correlated with items 4, 12, 13 and 32. Items 4, 12 and 13 were negatively correlated (r = -.352-.368, p = .024-.018). Item 32 was positively correlated (r = .338, p = .031). These items described individuals who had difficulty knowing what they were feeling, didn't visualize spatial concepts, found it difficult to imagine nonverbal thought and found it hard to visualize faces.

Next, I performed a factor analysis (excluding item 22, which was answered true by all participants). The scree suggested 3 factors, which I extracted. Many items significantly loaded on multiple factors, but I assigned each to the factor it loaded most strongly on. I then derived scores by reverse-scoring negatively loaded items and summing all items together.

Factor 1 included, in order of strength of correlation, items 7 (.807), 12 (.780), 25 (.738), 5 (-.662), 38 (-.656), 15 (.619), 2 (.598), 31 (.511), 39 (-.494), 13 (.401), 6 (-.353) and 11 (.336).

In general, factor 1 described visuo-spatial thinkers, who thought in spatial relationships represented visually, could mentally rotate objects, could easily imagine how someone's thought process could be entirely nonverbal, and so forth. The negatively correlated items mostly related to verbal thinking, such as having an internal monologue, using a phonological loop to boost verbal memory, and talking to plan a task. However, an item representing nonsensory thought (6) was also negatively correlated. In addition, they felt like part of their brain would turn off when they looked at a face (31), which may represent sensory overload or else prosopagnosia.

The average score on factor 1 was 14.19+/-6.045, with a range of 0-24 (possible range 0-24). This factor was positively correlated with AQ communication (r = .391, p = .013), suggesting that high scorers on this factor had more difficulty with social communication. It was also negatively correlated with age (r = -.322, p = .040).

Factor 2 included items 24 (.670), 20 (.632), 10 (.620), 27 (.601), 16 (.590), 9 (.541), 33 (.497), 17 (.470) and 8 (.348).

This factor also included some items related to visual thinking, such as visualizing a low-imagery sentence and visualizing writing when listening to speech. However, it also included mental representations from other modalities. In general, this factor seemed to represent vivid mental links between sensory cues and internal experiences, such as thought cued by location, hearing a person's voice when remembering their words, feeling as if they're doing an action when they watch someone else do it, feeling happy when they see or hear a word related to their interests and feeling like objects have personalities.

The average score on factor 2 was 11.02+/-4.793, with a range from 0-18 (possible range 0-18). Factor 2 and factor 1 were positively correlated (r = .366, p = .016). In addition, factor 2 was positively correlated with AQ social (r = .371, p = .020) and AQ attention to detail (r = .409, p = .010), suggesting that high scorers on factor 2 were less sociable and more likely to notice small details in their environment. This factor was also positively correlated with FrSBe total (r = .462, p = .003), disinhibition (r = .333, p = .033) and executive dysfunction (r = .521, p = .001), suggesting that high scorers tended to have trouble with planning, organization and inhibiting impulses.

Factor 2 was also positively correlated with Reaction to Eye Contact(REC) factor 1, indicating that high scorers on this factor tend to find eye contact unpleasant and pointless, but feel pressured to make eye contact anyway.

Factor 3 included items 30 (-.681), 4 (.588), 26 (.569), 34 (-.575), 32 (-.556), 23 (.494), 35 (.482), 14 (.459), 28 (-.395), 29 (.377) and 19 (.330).

This factor included several items related to verbal thinking, such as not having concepts they can't verbalize, talking themselves through visuospatial tasks and being able to imagine how thought could be entirely verbal, as well as not being able to imagine more than 3 spatial dimensions. However, there were also a number of items related to good self-awareness, such as knowing their own feelings, noticing hunger while concentrating and having an accurate mental representation of their own face. In addition, they didn't find visualizing faces difficult, had a good time sense that was not thrown off by changes in location, and thought of numbers as a line going from smaller to larger.

The average score on factor 3 was 8.49+/-5.239, with a range of 1-22 (possible range 0-22). Factor 3 was negatively correlated with AQ communication (r = -.506, p = .001), suggesting that high scorers on this factor tended to have better social communication skills. It was also negatively correlated with FrSBe total (r = -.336, p =.039), suggesting that these individuals tended to have better executive functions. In addition, this factor was negatively correlated with TAS total score (r = -.430, p = .006), difficulty describing emotions (r = -.586, p < .001) and external orientation (r = -.480, p = .001), suggesting that high scorers on this factor found it easy to describe their emotions and thought that emotional experiences were important.

This factor was negatively correlated with REC factors 1 (r = -.732, p < .001) and 2 (r = -.620, p = .032) and positively correlated with REC factor 3 (r = .418, p = .030), suggesting that high scorers on this factor tended to make a lot of eye contact and find eye contact beneficial to social processing, and did not have strong negative reactions to eye contact.

Lastly, 6 items didn't load on any factor - difficulty believing others could have different likes and dislikes (3), letter-color synesthesia (18), similar emotional attachment to humans and animals (21), verbal interference in math (36) and task shifting (37) and having an easier time thinking if moving (40).

Appendix: Autism Internal Experience items and responses:
ItemsTrueFalseNot sure
1. I have to consciously translate my thoughts into words.
31
72.1%
10
23.3%
2
4.7%
Total: 43
2. In my mind, concepts are usually represented nonverbally.
25
58.1%
10
23.3%
8
18.6%
Total: 43
3. I find it hard to believe that others could have different likes and dislikes from my own.
12
27.9%
28
65.1%
3
7.0%
Total: 43
4. I almost always know how I'm feeling.
8
18.6%
33
76.7%
2
4.7%
Total: 43
5. I think in a running stream of words with few or no images.
11
25.6%
31
72.1%
1
2.3%
Total: 43
6. My mental representation often has no sensory component whatsoever.
11
25.6%
18
41.9%
14
32.6%
Total: 43
7. I think in spatial relationships.
20
46.5%
13
30.2%
10
23.3%
Total: 43
8. When I am hearing someone talk, I visualize written text of what they are saying.
12
27.9%
28
65.1%
3
7.0%
Total: 43
9. When I am reading written words, I imagine that I can hear the words spoken.
30
69.8%
12
27.9%
1
2.3%
Total: 43
10. If I remember what someone said to me, I literally hear their words in my head.
35
81.4%
5
11.6%
3
7.0%
Total: 43
11. When I think of a concept such as 'dog', I see a bunch of images of dogs I've met or seen.
23
53.5%
16
37.2%
4
9.3%
Total: 43
12. When I think of the spatial relationship between two things, I represent this as a visual image.
33
76.7%
6
14.0%
4
9.3%
Total: 43
13. I find it easy to imagine how someone's conscious thought process could be entirely nonverbal.
18
41.9%
18
41.9%
7
16.3%
Total: 43
14. I find it easy to imagine how someone's conscious thought process could be exclusively verbal.
12
27.9%
23
53.5%
8
18.6%
Total: 43
15. When I hear or read a sentence like 'the number 8 when rotated 90 degrees looks like a pair of glasses', a visual image appears in my head as I process it.
37
86.0%
4
9.3%
2
4.7%
Total: 43
16. When I hear or read a sentence like 'addition, subtraction and multiplication are all math skills', a visual image appears in my head as I process it.
21
48.8%
19
44.2%
3
7.0%
Total: 43
17. Objects often seem like they have personalities to me.
27
62.8%
15
34.9%
1
2.3%
Total: 43
18. When I see certain letters, I automatically perceive them as having certain colors.
8
18.6%
31
72.1%
4
9.3%
Total: 43
19. I can easily mentally arrange events on a timeline based on when they happened, even if they happened in different locations.
20
46.5%
13
30.2%
10
23.3%
Total: 43
20. I find when I go to a place I've gone before, I find myself thinking about whatever I was thinking about when I was last there.
22
51.2%
15
34.9%
6
14.0%
Total: 43
21. I feel a similar degree of emotional attachment to both humans and animals that are close to me (eg loving a pet as a family member).
33
76.7%
7
16.3%
3
7.0%
Total: 43
22. If I don't know what to expect in a situation, it makes me uneasy.
43
100.0%
0
0.0%
0
0.0%
Total: 43
23. Without external clues, I can make a reasonable guess as to what time of day it is (within a half an hour of the true time).
12
27.9%
24
55.8%
7
16.3%
Total: 43
24. I tend to think of events in different locations as if they occurred in different timelines, for example if taking a class that I don't take every day, while in class I feel like the last class day was yesterday even though it was actually several days ago.
23
53.5%
12
27.9%
8
18.6%
Total: 43
25. I can visualize an object and then mentally rotate it to look at every side of it.
29
67.4%
10
23.3%
4
9.3%
Total: 43
26. Even when I'm concentrating on something else, I always notice when I'm starting to get hungry.
8
18.6%
33
76.7%
2
4.7%
Total: 43
27. When I hear or read a word related to one of my interests, I feel a warm, happy feeling.
32
74.4%
8
18.6%
3
7.0%
Total: 43
28. I can imagine what 4 or 5 dimensional space might look like.
10
23.3%
25
58.1%
8
18.6%
Total: 43
29. I think of numbers as a line going from smaller to larger.
16
37.2%
21
48.8%
6
14.0%
Total: 43
30. I know a lot of concepts that are difficult or impossible to describe verbally.
37
86.0%
3
7.0%
3
7.0%
Total: 43
31. When I look at a person's face, I feel like part of my brain is turning off.
21
48.8%
16
37.2%
6
14.0%
Total: 43
32. I find it more difficult to visualise a person's face than to visualise a non-living object of similar complexity and familiarity to me.
31
72.1%
9
20.9%
3
7.0%
Total: 43
33. When I watch someone else do an motor activity (especially one I've done in the past), it feels a bit like I'm doing that activity myself.
19
44.2%
20
46.5%
4
9.3%
Total: 43
34. When I look at my reflection, my face doesn't look the way I expected it to look.
29
67.4%
9
20.9%
5
11.6%
Total: 43
35. When I try to do a visual and/or spatial task, I 'talk myself through it', either overtly or silently.
20
46.5%
18
41.9%
5
11.6%
Total: 43
36. I'd find it harder to solve a math problem while repeating a word (eg 'Monday') over and over to myself.
32
74.4%
4
9.3%
7
16.3%
Total: 43
37. I'd find it harder to shift between two tasks in a pattern (eg 'sort by shape' versus 'sort by color') if I was also required to repeat a word (eg 'Monday') over and over to myself.
29
67.4%
5
11.6%
9
20.9%
Total: 43
38. When I'm trying to remember a phone number long enough to dial it, I recite the phone number to myself over and over.
32
74.4%
10
23.3%
1
2.3%
Total: 43
39. I talk to myself when trying to plan out how to do a task.
32
74.4%
8
18.6%
3
7.0%
Total: 43
40. I find it easier to think if I'm moving (eg pacing).
24
55.8%
12
27.9%
7
16.3%
Total: 43