هنگامی که جبران خسارت شکست خورد: نقص توجه در پیری سالم ناشی از حواس پرتی های بصری

Abstract Age related changes in frontal lobe functions are often related to attentional deficits that lead to increased distractibility by irrelevant stimuli. However, attentional functions have been reported not to decline in general with increasing age but simply be too slow to deal properly with distraction in time. Therefore older people might be able to compensate for distraction quite efficiently with sufficient processing time. Compensation, however, might fail when early perceptual processing is affected by distraction already. In the present study, a change in luminance or in orientation had to be detected in a sequence of two visual frames. Older participants showed reduced performance only when luminance and orientation changes were presented simultaneously at separate locations (perceptual conflict condition). Sensory ERP components were not overall altered with increasing age. Only in conflicting trials, a strong bias towards physically more salient information was observed. Additionally, older adults showed markedly delayed ERP-correlates of fronto-central control mechanisms in the conflict condition. The data indicate that processing deceleration cannot compensate for perceptual conflicts induced by mis-weighting of incoming information.

Introduction With increasing age, attention, working memory, episodic memory and other cognitive functions change in efficiency (Craik & Salthouse, 2000). The most influential hypotheses about the origin of age-related cognitive alterations are the “processing speed hypothesis” in which a global decline in processing speed is assumed (=general slowing; Salthouse, 1996), and the “inhibitory deficit hypothesis” which proposes an inability to control interference from task-irrelevant information with increasing age as a core deficit affecting various cognitive abilities (Hasher and Zacks, 1988 and Kane et al., 1994). Although there is accumulating evidence that general slowing might not be as general as originally assumed, e.g., from studies that report differential slowing factors in different sets of tasks (Lima, Hale, & Myerson, 1991) or from findings of widely unaffected early components of the event-related potential of the EEG (=ERP) with increasing age (Yordanova, Kolev, Hohnsbein, & Falkenstein, 2004), some studies show substantial contribution of slowing to behavior (Bugg et al., 2007 and Bugg et al., 2006). In a recent study, Gazzaley et al. (2008) investigated possible interactions between general slowing and inhibitory deficits in a working memory task. They found evidence for a direct interaction between alterations in neural processing speed and the ability to suppress irrelevant information at an early processing stage. Control processes that occur later in the time course of information processing appear to be less affected by age-related decline. Thus, Gazzaley et al., 2008 claim that suppression ability is not abolished with normal aging, but rather delayed to a later processing stage (see also Andrés, Parmentier, & Escera, 2006). Such a pattern of behavior can also be observed in attentional tasks. Slowing in the ability to allocate attention to a position where a non-informative cue was presented preceding the target stimulus, the so-called inhibition of return (=IOR; for a review see Klein, 2000, Posner and Cohen, 1984 and Posner et al., 1985), is delayed, but not abolished with increasing age (Castel et al., 2003 and Wascher et al., 2011) although electrophysiological measures indicated qualitative changes in processing irrelevant information of the cue (Wascher et al., 2011). Thus, delayed responses in attentional tasks might be due to a relatively longer lasting capture of attentional resources with increasing age and not to functionally deficient attentional mechanisms in general (see also Lien, Gemperle, & Ruthruff, 2011). Hence, with increasing age, behavior might be slower but comparably as efficient with respect to accuracy as behavior of in young participants, as long as older adults have sufficient time to process. In fact, it has been reported that older people respond slower in order to be more accurate (Falkenstein et al., 2001, Hoffmann and Falkenstein, 2011 and Wild-Wall et al., 2008). This rather strategic approach can only succeed as long as the processing stream of incoming information is intact. Yet, this strategy should not be possible when signals arrive in fast succession or even simultaneously. Here distraction may be able to immediately distort the processing of relevant signals (Enns & Di Lollo, 2000). Such sensory distractibility might suffice to severely impair higher cognitive processing like working memory (Gazzaley et al., 2008). Sensory distractibility with increasing age due to salient signals is a disputed issue. Pratt and Bellomo (1999), reported stronger attentional capture in older, compared to young adults. Phillips and Takeda (2010) demonstrated an increase of phase locking in the gamma band for older adults, indicating stronger susceptibility to bottom-up mechanisms. Additionally, older adults show severe deficits in visual masking (Kline & Birren, 1975) as well as in change blindness tasks (Rizzo et al., 2008). Opposed to these findings, some evidence for intact distracter inhibition abilities in older adults were reported (Lien et al., 2011 and Wnuczko et al., 2012). Salient distracters in a distributed display do not necessarily lead to deficits in selecting intentionally focused stimulus features, not even with abrupt onsets (Lien et al., 2011). The source for this apparent contradiction might be the dynamic interplay between voluntary (top-down controlled) and stimulus driven (bottom-up) mechanisms in attentional selection. The biased competition account of selective attention (Desimone, 1998 and Desimone and Duncan, 1995) proposes that incoming signals compete with each other in sensory processing and only the information that has been most amplified (bottom-up and/or top-down) wins and reaches consciousness. Within this system, the saliency of a particular stimulus, the amount of its intentional amplification and the neural sensitivity to its information determine whether it is selected for further processing or not (Beste et al., 2011, Kastner and Ungerleider, 2001, Reynolds and Desimone, 2003 and Reynolds et al., 2000). Therefore, there are several possible sources for the impact an irrelevant distracter has on stimulus processing and not all appear to be impaired with increasing age. In the present study, a change detection task was used to test the distractibility to salient stimuli with increasing age in the framework of the biased competition account. Participants had to detect a luminance or orientation change of a laterally presented bar that could be accompanied by a change of the other dimension at the opposite location (Wascher & Beste, 2010b). In case of such a competition of incoming signals, participants often fail to see the luminance change. Performance in this task highly depends on the perceptual sensitivity for intended information (Beste et al., 2011), attentional effort paid to the task (Sänger & Wascher, 2011) and the relative salience of relevant signals (Wascher and Beste, 2010b). Most importantly, it has been demonstrated that the outcome of early perceptual processing that initially weight the competing stimuli is highly predictive for successful selection (Schneider, Beste, & Wascher, 2012). In order to uncover subtle age-related differences in processing of such types of stimuli, event-related potentials of the EEG (ERPs) were measured. By means of the ERPs, the studies that were mentioned above distinctively described cortical correlates of sub-processes assigned to initial sensory stimulus processing (posterior asymmetries over visual areas in the N1 range=N1pc1 ), the intention based selection of relevant information (N2pc) and executive control functions that are related to the evoked perceptual conflict (an anterior N2, most probably reflecting activation of the anterior cingulate cortex [ACC]). Age-related deficits in the intentional control of irrelevant and relevant signals should be visible in the N1pc already. Reduced abilities of older adults to amplify intended information should lead to lower amplitudes when such information is presented. This should account for all conditions in which a relevant change (either luminance or orientation) is presented. Moreover, in those conditions slowed sensory processing should be visible in the latency of the N1. In conflict trials, when irrelevant stimuli compete with relevant information, increased susceptibility to bottom-up information, due to reduced distracter control, should not only lead to selectively increased error rates in this particular condition, but also to a bias of attention towards more salient information for early processing stages as reflected in the N1pc. However, one might also hypothesize that the sensory portion of the attentional selection remains widely unaffected with increasing age. In accordance to previous findings of intact early sensory processing (Yordanova et al., 2004), the deficit might occur not earlier than in the selection of relevant information, indicated by unaltered components in the N1 range but reduced amplitudes for the N2pc, at least when conflicting information is presented. However, due to the frontal decline with increasing age, a conflict induced by contradicting information in this task may be simply resolved less efficiently. In this particular task, we have reported that more salient distracters induce larger conflicts and larger amplitudes and shorter latencies of the fronto-central N2 in conflict trials (Wascher & Beste, 2010b). Given that older participant have problems in activating this frontal control instances, one would expect higher error rates accompanied by a decrease of amplitude of the N2 while no change would be expected in the N1 range. All three hypothesized sources, distorted control of stimulus driven processes (bottom-up), impaired ability to intentionally reallocate attention and deficient conflict monitoring would fit into the concept of increased distractibility with increasing age. In all cases, behavioral effects in terms of reduced accuracy should be restricted to the conflict condition, since only in this condition relevant and irrelevant information have to be separated. Prolonged response times might occur in all conditions, since general slowing should also affect responses to singular events.

Results 4.1. Behavioral data 4.1.1. Accuracy Overall, older participants committed more errors than younger ones, F(1,22)=16.216, p=.001 (see Fig. 2, left panel). Error rates were enhanced when participants attended to luminance, F(1,22)=5.616, p=.027, and with lower contrast, F(1,22)=50.502, p<.001. The type of change significantly influenced the error rates, F(3,66)=57.832, ε=.632, p<.001. More errors were committed when the unattended dimension changed compared to changes of the attended dimension, F(1,22)=34.832, p<.001). Moreover, when both dimensions changed at one location the error rate also slightly increased compared to single changes, F(1,22)=9.256, p=.006. Most errors were committed in the conflict condition, F(1,22)=135.351, p<.001. This latter effect was more pronounced in older adults as demonstrated in the interaction of change by age groups, F(3,66)=6.976, p<.001. This interaction, however, was restricted to the comparison of changes of both features at one location and conflict trials, F(1,22)=14.375, p=.001. Error rates (left panel) and response times (right panel) for all conditions. ... Fig. 2. Error rates (left panel) and response times (right panel) for all conditions. Changes of the attended stimulus dimension (A), of the unattended stimulus dimension (U), of both dimensions at one location (LOU) and the conflict condition where changes of both dimensions were distributed across the two locations (LOB) are plotted always in the same order. Age related deficits in accuracy were visible only in conflict trials. Delayed responses across all conditions may reflect general slowing. Figure options 4.1.2. RTs As shown by Fig. 2 (right panel), older participants responded more slowly than younger participants, F(1,22)=28.525, p<.001. The overall analysis revealed that responses were faster when participants attended to orientation, F(1,22)=6.204, p=.021, and when the contrast was high, F(1,22)=54.571, p<.001. The latter effect was slightly larger for older adults, which was visible in the interaction of contrast by age, F(1,22)=3.465, p=.076. 4.2. EEG data 4.2.1. Occipital N1 4.2.1.1. Unilateral changes Considering N1 latency evoked by unilateral changes of only one dimension (see Fig. 3) neither any interaction with nor the main effect of age, F(1,22)=.886, p>.2, reached significance. Event-related potentials of the EEG (ERP) evoked by unilateral changes of only ... Fig. 3. Event-related potentials of the EEG (ERP) evoked by unilateral changes of only one stimulus dimension (collapsed for LUM and ORI) over parietal occipital sites (P7/8; depicted separately for contra and ipsilateral leads). As for all other ERP plots, data were collapsed for dark and bright stimuli. There is no obvious age effect at contralateral electrodes. However, a small ipsilateral positive deflection appears to be missing in older adults. Figure options Also for the amplitude of this component, no overall group effect for age was observed, F(1,22)=.945, p>.2. As evident from Fig. 3 (right panel), there is some alteration in the ERP of older adults primarily at leads ipsilateral to the location of the change. Thus, considering only the amplitudes at ipsilateral sites, a tendency towards an age effect was observed, F(1,22)=3.056, p=.094, indicating less ipsilateral positivity in the N1 window for older adults. 4.2.1.2. Conflict trials (bilateral changes) Also in conflict trials, N1 peaks were remarkable similar across the two age groups (see Fig. 4). For the peak latency, no effect including the factor age reached significance in conflict trials (see Fig. 4). Event-related potentials of the EEG (ERP) evoked in conflict trials parietal ... Fig. 4. Event-related potentials of the EEG (ERP) evoked in conflict trials parietal occipital sites (P7/8; depicted separately for contra and ipsilateral leads with respect to the luminance change). The most obvious effect is an increase of the N1 amplitude ipsilateral to the luminance change, i.e., contralateral to the orientation change, independently from the stimulus dimension that has to be attended. Figure options For N1 amplitude, old participants showed a bias towards orientation changes in particular for bright stimuli, F(1,22)=4.654, p=.042. Again, no overall group effect of age was observed, F(1,22)=.031, p>.2. 4.2.2. Event-related lateralizations 4.2.2.1. Unilateral changes The N1pc (of the ERLs; see Fig. 5, left panel) for unilateral changes peaked earlier for high contrast stimuli, F(1,22)=112.1, p<.001. This effect tended to be larger in older adults, F(1,22)=3.187, p=.088. Overall, the N1pc showed longer latencies in older adults, F(1,22)=14.777, p=.001, indicating some slowing in information processing. Event-related lateralizations (ERLs=contra-ipsi difference waves) over parietal ... Fig. 5. Event-related lateralizations (ERLs=contra-ipsi difference waves) over parietal occipital sites for unilateral changes (left panel) and conflict trials (right panel). For unilateral changes delayed peak latencies are visible for older adults. For the conflict trials regions of significant age effects are marked. In the time range of the N1, older adults appear to be attracted by the presumably more salient orientation change. Figure options Neither the main effect of age nor any interaction including the group factor age reached significance for the amplitude of the N1pc (all F’s<1). 4.2.2.2. Conflict trials (bilateral changes) Since there were no distinct peaks measurable for the conflict condition (see Fig. 5, right panel), we calculated mean amplitudes for the N1pc (170–210 ms) and N2pc (250–300 ms). In the N1 range, a stronger asymmetry towards orientation changes was visible in older adults, F(1,22)=4.133, p=.054, which replicates the results for the N1 peak amplitudes for this condition. This effect was more pronounced for bright stimuli, F(1,22)=5.694, p=.026. Only the conflict condition a subsequent N2pc was visible that indicated attentional re-allocation of attention towards the task relevant stimulus feature (see Wascher & Beste, 2010b). If luminance was the intended dimension, a delayed turnover towards the relevant dimension was visible in older adults. In case of orientation as the relevant dimension, the data showed smaller amplitudes in the time range of the N2pc for the older group. In other words, in the N2-range, older adults showed a stronger attentional orientation towards luminance changes, F(1,22)=4.000, p=.058. No other effect reached significance. 4.2.3. Fronto-central N2 4.2.3.1. Peak analyses The N2 peak in the difference waves (LOB–LOU; see Fig. 6, left lower panel) was reliably delayed in older adults, F(1,22)=39.800, p<.001. No interaction of the factor age with any experimental factor was observed. Event-related potentials of the EEG (ERP) from a fronto-central electrode (FCz). ... Fig. 6. Event-related potentials of the EEG (ERP) from a fronto-central electrode (FCz). Despite essential differences in the basic characteristics of the wave shapes between age groups (see upper panel), the difference waves between conflict trials and unilateral changes of both dimensions revealed strikingly similar fronto-central peaks (N2) for both age groups that only showed much more variance in peak latency in older adults. In the right lower panel the interrelation between N2 latency and accuracy in conflict trials is depicted. Figure options Despite the apparent reduction of the N2 amplitude in older adults (see Fig. 6), the age of participants did not affect the amplitude of the N2 (main effect of age, F(1,22)=1.491, p=.235). All decrease of the N2 amplitude in the grand averages of the difference waves derived from the higher temporal variability of this component in older adults (latency jitter). 4.2.3.2. Correlation analyses The apparent reduction of the frontal N2 amplitude in older adults as visible in the difference waves plotted in Fig. 6 was due to an enormous variance in peak latency within this age group. Since we have hypothesized that the speed of control processes may be predictive for the ability to recognize distracted signals, we calculated correlations between N2-latency and error rates for the conflict trials (see Fig. 6, right lower panel). Because of the small sample size and the different scales of the measures that were entered into the analysis, Spearman’s Rho with one-tailed significance level is reported. Across all subjects, a significant correlation was observed, ρ=.666, N=24, p<.001. Restricting the analysis to old participants, there was still a significant effect, ρ=.515, N=12, p=.043. As obvious from Fig. 6, this effect disappeared when young participants were tested alone, ρ=−.147, N=12, p>.2. Also a partial correlation with age as the controlling variable between N2 latency and error rates in the conflict trials reached significance, r(21)=.350, p=.051.