Biology subjects

Asplund, C. L.

Publications and source records attributed to Asplund, C. L..

3 recordsLinked to original sources

bioRxiv2023

A test of time-resolved functional MRI with subsecond event durations

The hemodynamic signal measured with functional neuroimaging (fMRI) is notoriously sluggish, making it challenging to draw inferences about brief mental events. Relative timing differences in the response, however, can distinguish between changes in the duration or intensity of underlying neural activity (Henson et al., 2002). Specifically, increases in stimulus duration are predicted to delay the response peak and increase its amplitude, whereas increases in stimulus intensity should affect only peak magnitude (Friston, 2005). Although these different relationships have been empirically demonstrated using stimulus durations of several seconds, it remains unclear whether similar effects can be reliably detected on the sub-second timescales typical of cognitive events. Here we tested these predictions using brief visual and auditory stimuli that varied in intensity and duration. In Experiment 1 (n=15), stimuli were presented at three durations (100, 300, and 900 ms) and three intensities in a slow event-related design with a rapid TR (625 ms). In Experiment 2 (n=14), the stimulus durations were extended (1000, 1200, and 1800 ms) to enhance signal strength. Fitting the observed fMRI signals to parameterised hemodynamic response functions revealed that changes in stimulus duration affected both peak latency and magnitude, whereas changes in intensity affected only peak magnitude. Reliable effects were observed for duration differences as small as 200 ms, particularly when the evoked responses were strong. These findings support the validity of existing hemodynamic models and identify conditions under which sub-second timing differences can be reliably detected with fMRI. They also support the use of time-resolved fMRI as a principled and practical tool for mental chronometry in cognitive neuroscience.

neuroscience↗

bioRxiv2022

There is no fundamental trade-off between prediction accuracy and feature importance reliability

There is significant interest in using neuroimaging data to predict behavior. The predictive models are often interpreted by the computation of feature importance, which quantifies the predictive relevance of an imaging feature. Tian and Zalesky (2021) suggest that feature importance estimates exhibit low split-half reliability, as well as a trade-off between prediction accuracy and feature importance reliability across parcellation resolutions. However, it is unclear whether the trade-off between prediction accuracy and feature importance reliability is universal. Here, we demonstrate that, with a sufficient sample size, feature importance (operationalized as Haufe-transformed weights) can achieve fair to excellent split-half reliability. With a sample size of 2600 participants, Haufe-transformed weights achieve average intra-class correlation coefficients of 0.75, 0.57 and 0.53 for cognitive, personality and mental health measures respectively. Haufe-transformed weights are much more reliable than original regression weights and univariate FC-behavior correlations. Original regression weights are not reliable even with 2600 participants. Intriguingly, feature importance reliability is strongly positively correlated with prediction accuracy across phenotypes. Within a particular behavioral domain, there is no clear relationship between prediction performance and feature importance reliability across regression models. Furthermore, we show mathematically that feature importance reliability is necessary, but not sufficient, for low feature importance error. In the case of linear models, lower feature importance error is mathematically related to lower prediction error. Therefore, higher feature importance reliability might yield lower feature importance error and higher prediction accuracy. Finally, we discuss how our theoretical results relate with the reliability of imaging features and behavioral measures. Overall, the current study provides empirical and theoretical insights into the relationship between prediction accuracy and feature importance reliability.

neuroscience↗

bioRxiv2019

Overlapping attentional networks yield divergent behavioral predictions across tasks: Neuromarkers for diffuse and focused attention?

Attention is a critical cognitive function, allowing humans to select, enhance, and sustain focus on information of behavioral relevance. Attention contains dissociable neural and psychological components. Nevertheless, some brain networks support multiple attentional functions. Connectome-based Predictive Models (CPM), which associate individual differences in task performance with functional connectivity patterns, provide a compelling example. A sustained attention network model (saCPM) successfully predicted performance for selective attention, inhibitory control, and reading recall tasks. Here we constructed a visual attentional blink (VAB) model (vabCPM), comparing its performance predictions and network edges associated with successful and unsuccessful behavior to the saCPMs. In the VAB, attention devoted to a target often causes a subsequent item to be missed. Although frequently attributed to attentional limitations, VAB deficits may attenuate when participants are distracted or deploy attention diffusely. Participants (n=73; 24 males) underwent fMRI while performing the VAB task and while resting. Outside the scanner, they completed other cognitive tasks over several days. A vabCPM constructed from these data successfully predicted VAB performance. Strikingly, the network edges that predicted better VAB performance (positive edges) predicted worse selective and sustained attention performance, and vice versa. Predictions from the saCPM mirrored these results, with the networks negative edges predicting better VAB performance. Furthermore, the vabCPMs positive edges significantly overlapped with the saCPMs negative edges, and vice versa. We conclude that these partially overlapping networks each have general attentional functions. They may indicate an individuals propensity to diffusely deploy attention, predicting better performance for some tasks and worse for others.\n\nSignificance statementA longstanding question in psychology and neuroscience is whether we have general capacities or domain-specific ones. For such general capacities, what is the common function? Here we addressed these questions using the attentional blink (AB) task and neuroimaging. Individuals searched for two items in a stream of distracting items; the second item was often missed when it closely followed the first. How often the second item was missed varied across individuals, which was reflected in attention networks. Curiously, the networks pattern of function that was good for the AB was bad for other tasks, and vice versa. We propose that these networks may represent not a general attentional ability, but rather the tendency to attend in a less focused manner.

neuroscience↗