Cross-decoding supramodal information in the human brain

Testing the generalization of neural representations

Multivariate analysis methods are widely used in neuroscience to investigate the presence and structure of neural representations. Representational similarities across time or contexts are often investigated using pattern generalization, e.g. by training and testing multivariate decoders in different contexts, or by comparable pattern-based encoding methods. It is however unclear what conclusions can be validly drawn on the underlying neural representations when significant pattern generalization is found in mass signals such as LFP, EEG, MEG, or fMRI. Using simulations, we show how signal mixing and dependencies between measurements can drive significant pattern generalization even though the true underlying representations are orthogonal. We suggest that, using an accurate estimate of the expected pattern generalization given identical representations, it is nonetheless possible to test meaningful hypotheses about the generalization of neural representations. We offer such an estimate of the expected magnitude of pattern generalization and demonstrate how this measure can be used to assess the similarity and differences of neural representations across time and contexts.

Auditory Brainstem Response in Autistic Children: Implications for Sensory Processing

Purpose

Autistic individuals frequently experience sensory processing difficulties. Such difficulties can significantly impact important functions and quality of life. We are only beginning to understand the neural mechanisms of atypical sensory processing. However, one established way to measure aspects of auditory function is the auditory brainstem response (ABR). While ABR has been primarily hypothesized thus far as a means of early detection/diagnosis in autism, it has the potential to aid in examining sensory processing in this population.

Method

Thus, we investigated standard ABR waveform characteristics in age-matched groups of autistic and typically developing children during various stimulus and intensity conditions. We also examined within ear waveform cross correlations and inter-aural cross correlations (IACC) to assess replicability and synchrony of participants' ABRs, which was a novel approach to ABR analysis in this population.

Results

We observed longer peak latencies (esp. wave III and V) and interpeak latencies in the autism and typically developing groups in different conditions. There were no statistically significant results in cross correlation or IACC.

Conclusions

These results suggest that brainstem auditory function may differ slightly, but is mostly similar, between autistic and typically developing children. We discuss these findings in terms of their implications for sensory processing and future utility.

Adaptive coding occurs in object categorization and may not be associated with schizotypal personality traits

Processing more likely inputs with higher sensitivity (adaptive coding) enables the brain to represent the large range of inputs coming in from the world. Healthy individuals high in schizotypy show reduced adaptive coding in the reward domain but it is an open question whether these deficits extend to non-motivational domains, such as object categorization. Here, we develop a novel variant of a classic task to test range adaptation for face/house categorization in healthy participants on the psychosis spectrum. In each trial of this task, participants decide whether a presented image is a face or a house. Images vary on a face-house continuum and appear in both wide and narrow range blocks. The wide range block includes most of the face-house continuum (2.50-97.5% face), while the narrow range blocks limit inputs to a smaller section of the continuum (27.5-72.5% face). Adaptive coding corresponds to better performance for the overlapping smaller section of the continuum in the narrow range than in the wide range block. We find that participants show efficient use of the range in this task, with more accurate responses in the overlapping section for the narrow range blocks relative to the wide range blocks. However, we find little evidence that range adaptation in our object categorization task is reduced in healthy individuals scoring high on schizotypy. Thus, reduced range adaptation may not be a domain-general feature of schizotypy.

Size and Quality of Drawings Made by Adults Under Visual and Haptic Control

The aim of this study was twofold. First, our objective was to test the influence of an object’s actual size (size rank) on the drawn size of the depicted object. We tested the canonical size effect (i.e., drawing objects larger in the physical world as larger) in four drawing conditions — two perceptual conditions (blindfolded or sighted) crossed with two materials (paper or special foil for producing embossed drawings). Second, we investigated whether drawing quality (we analysed both the local and global criteria of quality) depends on drawing conditions. We predicted that drawing quality, unlike drawing size, would vary according to drawing conditions — namely, being higher when foil than paper was used for drawing production in the blindfolded condition. We tested these hypotheses with young adults who repeatedly drew eight different familiar objects (differentiated by size in the real world) in four drawing conditions. As expected, drawn size increased linearly with increasing size rank, whatever the drawing condition, thus replicating the canonical size effect and showing that this effect was not dependent on drawing conditions. In line with our hypothesis, in the blindfolded condition drawing quality was better when foil rather than paper was used, suggesting a benefit from haptic feedback on the trace produced. Besides, the quality of drawings produced was still higher in the sighted than the blindfolded condition. In conclusion, canonical size is present under different drawing conditions regardless of whether sight is involved or not, while perceptual control increases drawing quality in adults.

Blindfolded adults use mental transformation strategies for spatial scaling of tactile maps

The current study tested strategies of spatial scaling in the haptic domain. Blindfolded adults (N = 31, aged 20–24 years) were presented with an embossed graphic including a target and asked to encode a target location on this map, imagine this map at a given scale, and to localize a target at the same spot on an empty referent space. Maps varied in three different sizes whereas the referent space had a constant size, resulting in three different scaling factors (1:1, 1:2, 1:4). Participants’ response times and localization errors were measured. Analyses indicated that both response times and errors increased with higher scaling factors, suggesting the usage of mental transformation stratergies for spatial scaling. Overall, the present study provides a suitable, novel methodology to assess spatial scaling in the haptic domain.

Cognitive map formation through haptic and visual exploration of tactile city-like maps

In this study, we compared cognitive map formation of small-scale models of city-like environments presented in visual or tactile/haptic modalities. Previous research often addresses only a limited amount of cognitive map aspects. We wanted to combine several of these aspects to elucidate a more complete view. Therefore, we assessed different types of spatial information, and consider egocentric as well as allocentric perspectives. Furthermore, we compared haptic map learning with visual map learning. In total 18 sighted participants (9 in a haptic condition, 9 visuo-haptic) learned three tactile maps of city-like environments. The maps differed in complexity, and had five marked locations associated with unique items. Participants estimated distances between item pairs, rebuilt the map, recalled locations, and navigated two routes, after learning each map. All participants overall performed well on the spatial tasks. Interestingly, only on the complex maps, participants performed worse in the haptic condition than the visuo-haptic, suggesting no distinct advantage of vision on the simple map. These results support ideas of modality-independent representations of space. Although it is less clear on the more complex maps, our findings indicate that participants using only haptic or a combination of haptic and visual information both form a quite accurate cognitive map of a simple tactile city-like map.