International Journal of Biomedical Imaging

International Journal of Biomedical Imaging / 2012 / Article
Special Issue

MRI in Neurosciences

View this Special Issue

Research Article | Open Access

Volume 2012 |Article ID 579786 |

Thomas A. Christensen, Kyle R. Almryde, Lesley J. Fidler, Julie L. Lockwood, Sharon M. Antonucci, Elena Plante, "Modulating the Focus of Attention for Spoken Words at Encoding Affects Frontoparietal Activation for Incidental Verbal Memory", International Journal of Biomedical Imaging, vol. 2012, Article ID 579786, 11 pages, 2012.

Modulating the Focus of Attention for Spoken Words at Encoding Affects Frontoparietal Activation for Incidental Verbal Memory

Academic Editor: Ramin Parsey
Received15 Apr 2011
Revised20 Aug 2011
Accepted24 Aug 2011
Published22 Nov 2011


Attention is crucial for encoding information into memory, and current dual-process models seek to explain the roles of attention in both recollection memory and incidental-perceptual memory processes. The present study combined an incidental memory paradigm with event-related functional MRI to examine the effect of attention at encoding on the subsequent neural activation associated with unintended perceptual memory for spoken words. At encoding, we systematically varied attention levels as listeners heard a list of single English nouns. We then presented these words again in the context of a recognition task and assessed the effect of modulating attention at encoding on the BOLD responses to words that were either attended strongly, weakly, or not heard previously. MRI revealed activity in right-lateralized inferior parietal and prefrontal regions, and positive BOLD signals varied with the relative level of attention present at encoding. Temporal analysis of hemodynamic responses further showed that the time course of BOLD activity was modulated differentially by unintentionally encoded words compared to novel items. Our findings largely support current models of memory consolidation and retrieval, but they also provide fresh evidence for hemispheric differences and functional subdivisions in right frontoparietal attention networks that help shape auditory episodic recall.

1. Introduction

Attention is known to alter neural processing at multiple levels of both the peripheral and central nervous systems, and both auditory and visual attention have been conceptualized as operating in both “top-down” and “bottom-up” modes [17]. Top-down mechanisms reflect goal-based control in order to direct attention to particular targets or to sustain attention over time. In contrast, bottom-up mechanisms have traditionally been defined by the phenomenon of reflexive attentional orienting, as when attention is drawn without intent by highly salient sensory stimuli such as a sudden loud noise or flash of light. Recently, however, some investigators have more broadly considered bottom-up effects as relevant for any incoming stimuli, with the relative saliency of the stimulus influencing whether it is ultimately encoded into memory [8]. Two recent theoretical models address the question of what roles stimulus saliency might play in first successfully encoding information into memory and then later retrieving it. As discussed below, the “Embedded Processes” model and the “Attention-to-Memory” model, while similar, also highlight the potentially divergent roles that attention at encoding may play in later recall. Brain mapping studies based on these models have only started to identify the neural substrates that underlie these cognitive processes in different domains.

In real life situations, numerous bits of information co-occur in the environment and vie for attention. Given that memory is a limited capacity system [911], the information encoded into memory must be restricted to what is relevant. According to the “Embedded Processes” model, attention is central to this process. Information is first encoded into working memory through an active process known as “attentional scanning” [12]. This process first searches through a set of potential memory items, then selects the most salient items and brings them into the “focus of attention” [1214]. However, even though the attentional scanner is responsible for selecting the appropriate stimuli, its capacity is limited, and it therefore can maintain focus on only small amounts of information at a time. This also places limits on the capacity of working memory. Cowan and colleagues have speculated that regions in and around the temporal-parietal junction are key to focusing attention [14]. More recently, others have found that the majority of functional neuroimaging data on working memory is consistent with the tenets of this model [15], with ventral posterior (inferior parietal cortex and intraparietal sulcus) regions associated with attentional focus and working memory maintenance, and lateral prefrontal regions involved in the executive control of attention and further manipulation of information, once it is in working memory [16, 17]. If this interpretation is correct, then manipulations of attention at encoding should preferentially affect activation in parietal regions.

Another recent paradigm, the “Attention-to-Memory” model [8, 1820], shares functional features with the previous model, but it also emphasizes a distinction between the roles of ventral and dorsal parietal cortex in memory retrieval. Successful recognition memory for intentionally studied items is predicted by increased activation of dorsal posterior parietal cortex at stimulus encoding, whereas activation in ventral posterior parietal cortex at encoding predicts success in incidental memory and perceptual priming tests [21, 22]. This latter observation suggests a role for this ventral region not only in the bottom-up encoding of highly salient or unexpected sensory stimuli, but also in the successful retrieval of unintentionally encoded memories [21]. According to these prevailing theories, retrieval of incidental or perceptual memory representations requires a bottom-up shift in selective attention, whereas retrieval of intentional memories requires a top-down mechanism that reflects the individual’s conscious intent to focus attention on incoming signals [21, 22]. Importantly, however, the vast majority of studies exploring the intersection of attention and episodic memory—and thus the theories on which these studies are based—are heavily grounded in the use of visual paradigms [2330]. Even studies that focus specifically on verbal encoding have mostly employed reading rather than listening tasks [3137]. It is therefore necessary to examine the effects of attention on item retrieval in listening tasks in order to establish whether the dorsal/ventral distinction observed in parietal cortex is specific to visual stimuli, or instead reflects a more generalized organization in this area of the brain.

To examine these ideas in greater depth, it is necessary to learn more about the role of attention in establishing a selective focus on task-relevant sensory stimuli. Moreover, in order to understand how limited attentional resources are allocated, it is also important to study the encoding of items both within the focus of attention as well as those peripheral to it. Our goal in this study was to devise a method to modulate attention toward spoken words during a simple listening task. Based on earlier studies [38], we predicted that words not intentionally within the focus of attention would be encoded into memory to some degree. During the encoding phase of our study [39], participants initially heard a list of words belonging to two semantic classes (animals or foods), and these words were presented by two different talkers (female or male). Participants were given explicit instructions to remember only the animal words presented in the female voice (the targets). Importantly, this instruction resulted in a focus on the female voice, which had the consequence of placing the nontarget food words presented in that voice also within the focus of attention. We predicted that these words would also carry higher salience at encoding than words heard in the male voice, creating what we refer to as the High-Attention condition in this study. Following this rationale, the food words presented in the male voice would not be fully in the primary focus of attention (the Low-Attention condition). We contrasted these two encoding conditions with responses to novel food words that were not previously heard at any time during the experiment. We reasoned that processing of these NEW words should reflect only the role of attention in initial word encoding, but not a later role in retrieval. Using this paradigm coupled with functional magnetic resonance imaging (fMRI), we were able to disambiguate the neural activation patterns associated with attention at first encoding from those reflecting the effects of attention at encoding on subsequent retrieval.

2. Materials and Methods

2.1. Participants

We recruited 14 native English-speaking adults (9 women; ages 18–49; mean 24 years) living in the Tucson area. Exclusion criteria were a history of speech, language, or other neurological disorders, and all volunteers reported good general health with no contraindications for MRI scanning. The study was approved by the University of Arizona Institutional Review board, and written informed consent was obtained from all participants. For one participant, behavioral response data are not available due to computer error.

2.2. Auditory Stimuli

Word stimuli were single, concrete nouns (one to three syllables) that were either names of animals or foods. Some words were recorded in an unfamiliar female voice, and some in an unfamiliar male voice. Stimulus durations ranged from 204–903 ms (mean = 532 ms) and were presented in an event-related format with an average interstimulus interval (ISI) of 871 ± 157 ms. ISI was jittered and “null” events were included to facilitate estimation of the hemodynamic responses for deconvolution analysis (see below). Ordering of words belonging to the different stimulus categories described below was pseudorandom, and participants listened to them through MRI-compatible stereo headphones (Resonance Technology Inc., Northridge, Calif, USA).

2.3. Procedure

The experimental design is shown in Figure 1. A prescan practice phase was followed by an encoding phase and a test phase (both with scanning). These were then followed by a surprise postscan memory test that specifically measured the listener’s capacity to remember any words that were encoded during the earlier phases of the experiment, whether or not the listener was asked to remember them. The results of the test phase and scan, along with results from the postscan behavioral test, are the subject of this report. Results obtained during the encoding phase have been published previously [39].

2.3.1. Encoding Phase and Scan

Participants listened to the animal and food words that were presented in either the male or female voice during a prescan practice period and during the encoding phase. Using a block design, the encoding phase consisted of two scans that were counterbalanced for order across participants. In one scan, approximately one-third of the total words used for this study were presented binaurally, with the male and female voices presented sequentially. In a second scan, two-thirds of the words were presented dichotically, with the female voice heard in one ear and the male voice heard in the other (order counterbalanced across participants). The ratio of targets (animal word + female voice) to nontargets was 1 : 3 in both tasks. Although all participants were explicitly instructed to focus on and remember the animal words presented in the female voice, these were not actually of interest for the purposes of this study. Rather, this instruction was used deliberately to assure that the primary focus of attention was on the female voice. As a result, food words presented in the female voice were associated with a higher level of attention (defined as the High-Attention condition) and food words presented in the male voice were associated with a lower level of attention (the Low-Attention condition).

2.3.2. Test Phase and Scan

An incidental memory scan immediately followed the encoding scan. During this scan, participants heard words presented by a second female speaker whose voice had not been heard earlier. This precluded the possibility that the voice in the memory task could provide a cue to the encoding context. Participants were asked to respond “yes” via button press if the word presented was one of their target words (animal words originally spoken in the female voice) and “no” to all other words. Therefore, both the decision (“not a target”) and response demands (selecting the “no” button) associated with all words analyzed in this study were identical. Participants were instructed to respond quickly but accurately. During this task, participants mostly heard nontarget words. A combination of High-Attention and Low-Attention food words (58 trials each) as well as 58 NEW food words that had not been presented previously were pseudorandomized and presented in a fixed order to all participants. To lessen fatigue, participants were tested over the course of two separate scans (each lasting ~9 min with a brief rest in between), during which a combined total of 232 words (randomized with 58 nulls) were presented. Data from the two scans were then concatenated for further analysis.

2.3.3. Surprise Postscan Memory Test

In order to assure that there was an effect of the attentional manipulation at encoding, we administered a surprise memory test within 15 minutes of the final scan. Participants were presented verbally with a list of 40 food items and asked to recall whether they had heard each word at any time during the experiment. The list consisted of 10 High-Attention, 10 Low-Attention, and 10 NEW food words that were all heard during the study phase and scan. A fourth category, 10 food items not presented in any of the scans, was added to measure false-alarm rate. These 40 words were presented verbally in a fixed pseudorandom order by the experimenter and the participants verbally responded “yes” or “no” to indicate whether they remembered having heard these words at any time during the experiment.

2.4. Whole-Brain fMRI Imaging

Scans were acquired with a 3.0T GE Signa VH/i scanner (General Electric Medical Systems, Madison, Wis, USA) equipped with a quad-head RF coil. First, T1-weighted, fast-spin echo (FSE) axial images covering the entire brain were acquired in 26 slices with an in-plane resolution of  mm. Next, two functional T2*-weighted scans were acquired using a spiral in/out pulse sequence (TR = 2.3 s, TE = 30 ms, flip angle = 90°, 26 slices at 5 mm with no gap, matrix = 64 × 64, FOV = 22 cm2) [40]. Finally, high-resolution spoiled gradient-echo (SPGR) images were obtained in the sagittal plane (TR = 30, TE = min, flip angle = 30°, 124 slices at 1.5 mm, matrix = 256 × 256, FOV = 25 cm2) and aligned with the FSE images for improved regional localization and coregistration of functional data across participants after transforming the images into Talairach space [41]. Due to scanner-related problems, we were unable to use the SPGR images from three participants, and instead relied on the first set of axial FSE anatomical images for alignment with the functional scans.

2.4.1. Data Analysis

Structural and functional brain images were analyzed for each participant individually with AFNI [42] followed by a group analysis. Blood-oxygenation-level-dependent (BOLD) contrast images were coregistered with anatomical data after preprocessing using standard procedures for slice-time correction, removing linear signal drift and correcting for head motion. All volumes were realigned to the base volume and spatially smoothed using a 6 mm Gaussian kernel. Data were then normalized to a scale of 0–100%, and functional images were coregistered to the structural data followed by transformation into standard Talairach space. The first four volumes in each run were also discarded to allow for T1 equilibration, and the two runs were then concatenated for further analysis. A general linear model using a gamma-spline hemodynamic response function was used to estimate magnitude parameters for events of interest for each stimulus condition in each individual. In order to capture the amplitude of the BOLD activity over time, eight separate activity models were developed. The first was time locked to the stimulus onset, and each of the seven subsequent models was offset by one-TR (2.3 s) increments from stimulus onset. Stimulus functions were then convolved with the fMRI time-series data from each individual. Parameter estimates for the resulting regressors for each condition were calculated using the least-squares fit of the models to the time-series data. Finally, a group analysis was performed with repeated-measures ANOVA (treating individuals as a random effect) to help confirm key cortical regions that showed differential neural activity for each of the four conditions. Monte Carlo simulation (8 mm FWHM blur; 1000 iterations) was used to correct the group data for multiple comparisons. Voxel-wise (uncorrected) threshold was , and minimum corrected cluster volumes in original space were 32 contiguous voxels at .

2.4.2. Data-Driven Cluster Analysis

To identify regions of interest (ROI) associated with positive BOLD signal across the High-Attention, Low-Attention, and NEW categories, we first combined the three group-averaged datasets. This allowed us to develop ROI that were associated with activation in any of the three conditions of interest. We were particularly interested in areas in and around the dorsolateral prefrontal cortex and the temporoparietal junction due to their proposed involvement in executive control as well as phonological short-term storage and language processing [12, 43, 44]. However, since the proposed involvement of these regions is based almost entirely on visual rather than auditory paradigms, and there is growing evidence for bilateral involvement in language processing [45, 46], we examined the following ROI in both hemispheres. Magnitude estimates and peak Talairach coordinates are listed in Table 1.

Left hemisphere Right hemisphere
, , , ,

Condition 1: High Attention
 Frontal lobe
  Medial frontal gyrus (6/32)−2, 3, 488.971, 12, 457.31
  Middle frontal gyrus (9/46)−52, 22, 283.2647, 18, 373.17
  Inferior frontal gyrus (47)−52, 17, −33.7246, 16, −44.39
  Insula (13)−36, 21, 54.9236, 21, 66.50
 Parietal lobe
  Supramarginal gyrus (40)−35, −62, 39(1)3.7639, −55, 46(2)3.39
 Superior temporal lobe
  Heschl’s gyrus (41/42)−50, −18, 105.5453, −17, 109.09
 Caudate body−14, −6, 194.3214, −6, 193.77
 Cerebellum—anterior lobe−32, −50, −273.8426, −53, −265.59
 Cerebellum—posterior lobe−7, −69, −232.816, −69, −233.04
 Thalamus—anterior−7, −13, 15.621, −9, 144.19

Condition 2: Low Attention
 Frontal lobe
  Medial frontal gyrus (6/32)−2, 3, 477.051, 12, 455.99
  Middle frontal gyrus (9/46)−50, 17, 273.7853, 16, 282.95
  Inferior frontal gyrus (47)−50, 24, −13.9849, 26, −23.11
  Insula (13)−34, 23, 44.3840, 16, 04.46
 Superior temporal lobe
  Heschl’s gyrus (41/42)−57, −21, 133.6051, −18, 108.45
 Cerebellum—anterior lobe−36, −47, −275.8931, −50, −273.69
 Cerebellum—posterior lobe−1, −70, −218.577, −70, −232.88
 Thalamus—anterior−8, −9, 94.825, −11, 125.21

Condition 3: NEW words
 Frontal lobe
  Medial frontal gyrus (6/32)−1, 9, 465.941, 12, 457.31
  Middle frontal gyrus (9)−52, 19, 273.2657, 22, 273.17
  Inferior frontal gyrus (45)−59, 22, 73.4849, 26, 03.18
  Inferior frontal gyrus (47)—, —, —51, 22, 12.89
  Insula (13)−42, 15, 03.1942, 14, −16.50
 Parietal lobe
  Angular gyrus/middle temp. gyr. (39)−41, −63, 243.7646, −65, 213.39
 Superior temporal lobe
  Heschl’s gyrus (41/42)−58, −18, 115.5452, −14, 109.09
 Cerebellum—anterior lobe−36, −54, −283.2138, −54, −273.00
 Cerebellum—posterior lobe−1, −69, −222.835, −72, −222.30
 Thalamus—anterior−8, −8, 124.703, −8, 113.25

Group-averaged data showing peak intensity values ( ) and coordinates (mm in Talairach space) for activated clusters (closest Brodmann areas in parentheses) in each of the three conditions compared to the resting state. Corrected activation threshold = , . (1)Activity spread into intraparietal sulcus in 58% of listeners. (2)Activation in SMG, but peak was more localized to angular gyrus in 29%, and intraparietal sulcus in 21% of listeners.

Masks corresponding to distinct anatomical regions of interest were developed, and ROI were defined based on the threshold-corrected regions of significant activation. Brain regions showing significant positive BOLD activation are listed in Table 1 and described as follows (approximate Brodmann areas in parentheses): Frontal Lobe (dorsal to ventral): medial frontal gyrus (6/32), middle frontal gyrus (9/46), inferior frontal gyrus (47/pars orbitalis), and anterior insular cortex (13). Parietal Lobe (inferior parietal lobule): angular gyrus (39) and supramarginal gyrus (40). We looked for additional ROI along the anterior-posterior axis of the superior temporal lobe due to its routine involvement in speech and language tasks [4751], but found only one with significant activation: superior temporal gyrus including and extending around the transverse temporal gyrus (BA41/42). Sub-cortical ROI included basal ganglia (caudate body) and thalamic nuclei (anterior). Finally, several ROI were localized to the anterior and posterior lobes of the cerebellum.

3. Results and Discussion

3.1. Surprise Postscan Memory Test

The surprise memory test was intended to provide proof of concept that the attentional manipulation at encoding actually produced an effect. Specifically, High-Attention words should be better remembered than Low-Attention words, and each of these should be better attended than NEW words. Conversely, participants should not indicate recognition for words introduced as foils that were not presented earlier at any time in the experiment.

Figure 1 shows the result of the group memory analysis (full variance model: , listeners; , two-tailed). As predicted, the strongest memory scores were for responses to the High-Attention words. These words showed the best hit rate at 93.8 ± 2.4% (mean ± SEM), compared to 70.8 ± 4.6% for Low-Attention words, and 51.0 ± 6.5% for NEW words (all means significantly different based on post-hoc, two-tailed ANOVA; see Figure 1). Additionally, words introduced as foils during the postscan memory test were correctly rejected 86.2 ± 3.3% of the time, indicating a low false alarm rate.

3.2. fMRI Results

A temporal analysis of the BOLD responses for each stimulus condition was performed by modeling the fMRI time series over a range of time lags, as shown in Figure 2. Although this analysis lacks the temporal precision of electrophysiological methods [1], it can provide valuable information about the relative timing of neural events associated with each test condition [31]. As stated above, eight separate hemodynamic response models were constructed for each stimulus type, and the peak of each model was time-shifted by one TR in order to capture peak responses across an 8-TR (18.4 s) time window. We then calculated the BOLD responses from the peak activation at each time lag, as shown in Figure 2. As expected, strong BOLD activity was observed in the superior temporal lobe in the vicinity of primary auditory cortex, and this activity did not vary with the three stimulus conditions (Table 1). In addition to the expected early activity in left superior temporal gyrus, BOLD responses associated with all three stimulus categories were, on average, 20% greater in right superior temporal gyrus, which is consistent with the other right-lateralized activation patterns discussed below.

3.2.1. Segregation of Function in Dorsal and Ventral Prefrontal Cortex

As shown in Figure 2(a), identification of words in all three categories was associated with activity in left middle frontal gyrus (a portion of dorsolateral prefrontal cortex—DLPFC) and this activity was closely mirrored in anterior insular cortex (Figure 2(b)). Furthermore, the time course of the BOLD response in these two left-hemisphere regions was similar for all three word categories regardless of the initial encoding condition. In contrast, right DLPFC showed a distinctly different response pattern, with High-Attention words yielding significantly greater responses relative to Low-Attention words, and the latter yielding significantly greater responses relative to NEW words. This pattern indicates a strong differential effect of attention at encoding in right DLPFC (BA9/46) compared to left DLPFC (Figure 2(a)) and bilateral insula (Figure 2(b)). Importantly, this result is consistent with earlier studies using written words that linked activity in this region of right DLPFC with post-retrieval processing and/or monitoring functions [37, 5254]. If right DLPFC is indeed involved selectively in post-retrieval processing, we would also expect, in accord with the “Attention-to-Memory” model, that words strongly attended at encoding might evoke a greater memory response than words only weakly attended at encoding. This graded pattern of activity was, in fact, observed for the High- and Low-Attention words, as shown in Figure 2(a). Conversely, this region would not be expected to activate in response to NEW words because these words were not yet encoded into memory. In accord with this hypothesis, only negligible positive BOLD activity was observed in right DLPFC in response to NEW words (Figure 2(a), right column).

As shown in Figure 2(c), another bilateral set of clusters was located in the inferior frontal gyrus (IFG). This area is believed to be an important rostral component of the so-called “ventral frontoparietal stream” that is frequently observed in studies involving the detection of novel or low-frequency events, particularly when they are unexpected [3]. Although this pathway has been observed in numerous visuospatial attention studies, our findings using auditory language stimuli are also consistent with this anatomical framework. Figure 2 shows how the three stimulus conditions in our study differentially activated the ventral frontoparietal stream. The focal point for these clusters in the frontal lobes was found in pars orbitalis (BA47), and in accord with the ventral frontoparietal model, activity was also lateralized to the right hemisphere (Figure 2(c), right column). Both the High- and Low-Attention conditions were associated with comparable activation in left and right IFG. However, in left IFG, a BOLD response on a similar time course was absent for the NEW words, indicating that the early response in IFG to previously encoded words may specifically reflect memory for these previously presented items. Despite the absence of an early BOLD response in left IFG for NEW words, these items were instead associated with a late-onset response in the right hemisphere (Figure 2(c), right column). In other words, responses to the NEW words differed from the previously attended words both in terms of hemispheric lateralization and in the late time course of activation in IFG (BA47), suggesting that this late activation more likely reflects the initial encoding of NEW words into memory.

3.2.2. Segregation of Function in Ventral Posterior Cortex

With respect to language processing, converging evidence suggests a key role for posterior parietal and temporal regions in verbal memory and attention [12, 31, 37, 44]. One memory model predicts that items correctly identified as previously encountered (old) will trigger increased activity in left intraparietal cortex, relative to missed old and correctly rejected novel items [31]. However, as noted above, selective attention is another cognitive function that is commonly associated with posterior parietal cortex, and it remains unclear how the influence of attention may affect the neural networks underlying recollection and familiarity, especially in the auditory domain.

It has been proposed that inferior parietal lobule, a region within ventral posterior parietal cortex, may serve a specialized function in the expression of attention, as proposed in both the Embedded-Processes model [14, 15] and the Attention-to-Memory (AtoM) model [8, 1820]. These models allowed us to generate informed predictions about the neural substrates underlying a possible attention-dependent memory effect in our study. If ventral posterior parietal cortex serves specifically as the “focus of attention” as in the Embedded-Processes model, we would expect activation in this region to be differentiated according to the level of attention directed toward each stimulus at encoding, as shown in Figure 1. Another prediction is that any region responsible for the reactivation of focused attention prior to memory retrieval should display activity earlier in the hemodynamic response than regions responsible for post-stimulus processing steps, such as semantic analysis.

As shown in Figure 2(d) and Table 1, the principal locus of activation in posterior parietal cortex was found bilaterally in the supramarginal gyrus (SMG). This activity, furthermore, was associated specifically with listening to and identifying the High-Attention words, but not the Low-Attention words. This is consistent with the notion that activation in and around the SMG is related to attentional scanning that placed our High-Attention words into the “focus of attention” according to the Embedded-Processes model. To examine the spatial location of the group activation in greater detail, the activation was overlaid onto the individual anatomical scans of each participant. This participant-by-participant anatomical localization revealed that the peak activation was located in the left SMG for 86% (12 of 14 individuals), and in the right SMG for 64% of the participants. As shown in Table 2, activation in the left hemisphere was completely isolated to the SMG in 36% of listeners, whereas in 50%, the activity also spread from the SMG into the adjoining intraparietal sulcus (IPS). In contrast, the primary focus of activation in the right hemisphere was more variable, localizing to the SMG for 50% of the participants, to the angular gyrus (AG) for 29% of participants, and to the IPS for 21% of participants. This greater variation likely reflects the greater variability of the anatomical landmarks within the posterior right hemisphere for our participants.

Left hemisphere Right hemisphere
Primary focusSecondary spreadPrimary focusSecondary spread


AG: angular gyrus (BA39); IPS: intraparietal sulcus (BA40/7); SMG: supramarginal gyrus (BA40); SPL: superior parietal lobule (BA7).

In contrast to these spatial activation patterns for previously encoded High-Attention words, the principal locus of activation in posterior parietal cortex for NEW words was associated with delayed activity in BA39; specifically AG and the adjacent portion of posterior middle temporal gyrus (pMTG) in both hemispheres (Figure 2(e); Table 3). In the left parietal lobe, the activity was centered within the AG for 64% of the participants (9 of 14 individuals) with additional activity that spreads around the peak activation into pMTG. In the right parietal lobe, the primary focus was centered more often in pMTG (64%) than in AG (36%) (Table 3). Thus the BOLD activation associated with NEW words was distinctly more ventral and posterior than the parietal activation pattern for previously encoded High-Attention words. Moreover, the time course of NEW word-related activity in right posterior parietal cortex was similar to that observed in right IFG for NEW words (compare Figures 2(c) and 2(e)). This finding suggests that the timing of activation in right BA39 may be an important determinant in how different stimulus contexts are represented in this region. The timing difference may reflect two distinct functional roles for this component of the frontoparietal attention network in processing previously encoded (High-Attention) words and newly encoded (NEW) words. One possibility is that the delayed response in BA39 may reflect the increased time required to access semantic content associated with these NEW words [55], or alternatively, activation may be associated more directly with processing novel events [3].

Left hemisphere Right hemisphere
Primary focusSecondary spreadPrimary focusSecondary spread


AG: angular gyrus (BA39); pMTG: posterior middle temporal gyrus (BA39); SMG: supramarginal gyrus (BA40).

Based on these results, we propose a neuroanatomical framework that involves two subnetworks in the ventral frontoparietal attention stream, as shown in Figure 3. The relatively early onset activation for previously encoded words in BA40 and BA47 suggests an early attentional role for this subnetwork (Figure 3, red pathway), while the delayed time course of NEW word responses in BA39 and BA47 could reflect the activity of a separate ventral pathway associated principally with bottom-up processing of novel word stimuli (green pathway). Our results are therefore consistent with previous findings that ventral parietal cortex plays a pivotal role in language-based tasks [37, 5557]. Inferior parietal activity has been found more reliably in studies of working memory than in those of explicit recall [15, 5860], and our results are also consistent with these findings. Moreover, during the encoding phase of our study (Figure 1), right BA40 was activated at the time our spoken stimulus words were first encoded [39]. The results of the memory phase in the present study show that right BA40 was reactivated at retrieval, a finding strongly in line with the transfer-appropriate processing model [10, 25, 2931, 6163]. We propose that early selective activation of right BA40 serves to initiate incidental memory for the studied words by bringing them back into the “focus of attention” that was established earlier during encoding [13, 6466].

4. Conclusion

In summary, our data support a functional framework in which the brain regions that are engaged in identifying words are also sensitive to the level of attention at the time the words were initially encoded. A structural model generated from our findings is illustrated in Figure 3. Our data are consistent with a “focus of attention” centered in right SMG (BA40), but we also identified a frontal region in right DLPFC (BA9/46) that is consistent with an “Attention-to-Memory” function in postretrieval processing. Other frontal regions (left DLPFC) were insensitive to the attentional manipulation at encoding, responding similarly to both the two studied word categories as well as to the newly encoded (NEW) words. The pattern in left DLPFC is therefore consistent with other cognitive processes such as executive control of attention that would not be expected to vary across our word categories. A key result was that once a spoken word was unintentionally encoded, subsequent retrieval of that word varied as a function of the initial level of attention directed toward the word at the time of encoding. This is consistent with the notion that attention is critical not only for efficient encoding of words into memory, but it also helps to preserve salient information that is required for the successful retrieval of that information from memory.


This research was supported by a grant from the NIH/National Institute on Deafness and Other Communication Disorders (K01 DC008812) to Thomas A. Christensen.


  1. M. H. Giard, A. Fort, Y. Mouchetant-Rostaing, and J. Pernier, “Neurophysiological mechanisms of auditory selective attention in humans,” Frontiers in Bioscience, vol. 5, pp. D84–D94, 2000. View at: Google Scholar
  2. C. D. Frith and K. J. Friston, “The role of the thalamus in “top down” modulation of attention to sound,” NeuroImage, vol. 4, no. 3, pp. 210–215, 1996. View at: Publisher Site | Google Scholar
  3. M. Corbetta and G. L. Shulman, “Control of goal-directed and stimulus-driven attention in the brain,” Nature Reviews Neuroscience, vol. 3, no. 3, pp. 201–215, 2002. View at: Google Scholar
  4. M. I. Posner and M. K. Rothbart, “Research on attention networks as a model for the integration of psychological science,” Annual Review of Psychology, vol. 58, pp. 1–23, 2007. View at: Publisher Site | Google Scholar
  5. A. Messinger, M. A. Lebedev, J. D. Kralik, and S. P. Wise, “Multitasking of attention and memory functions in the primate prefrontal cortex,” Journal of Neuroscience, vol. 29, no. 17, pp. 5640–5653, 2009. View at: Publisher Site | Google Scholar
  6. T. A. Polk, R. M. Drake, J. J. Jonides, M. R. Smith, and E. E. Smith, “Attention enhances the neural processing of relevant features and suppresses the processing of irrelevant features in humans: a functional magnetic resonance imaging study of the stroop task,” Journal of Neuroscience, vol. 28, no. 51, pp. 13786–13792, 2008. View at: Publisher Site | Google Scholar
  7. W. A. Freiwald, “Attention to objects made of features,” Trends in Cognitive Sciences, vol. 11, no. 11, pp. 453–454, 2007. View at: Publisher Site | Google Scholar
  8. R. Cabeza, E. Ciaramelli, I. R. Olson, and M. Moscovitch, “The parietal cortex and episodic memory: an attentional account,” Nature Reviews Neuroscience, vol. 9, no. 8, pp. 613–625, 2008. View at: Publisher Site | Google Scholar
  9. A. Baddeley, “The magic number and the episodic buffer,” Behavioral and Brain Sciences, vol. 24, no. 1, pp. 117–118, 2001. View at: Publisher Site | Google Scholar
  10. R. Desimone and J. Duncan, “Neural mechanisms of selective visual attention,” Annual Review of Neuroscience, vol. 18, pp. 193–222, 1995. View at: Google Scholar
  11. N. Lavie, “Distracted and confused?: selective attention under load,” Trends in Cognitive Sciences, vol. 9, no. 2, pp. 75–82, 2005. View at: Publisher Site | Google Scholar
  12. N. Cowan, W. Lichty, and T. R. Grove, “Properties of memory for unattended spoken syllables,” Journal of Experimental Psychology: Learning, Memory, and Cognition, vol. 16, no. 2, pp. 258–269, 1990. View at: Google Scholar
  13. N. Cowan, “Visual and auditory working memory capacity,” Trends in Cognitive Sciences, vol. 2, no. 3, pp. 77–78, 1998. View at: Publisher Site | Google Scholar
  14. N. Cowan, “An embedded-processes model of working memory,” in Models of Working Memory: Mechanisms of Active Maintenance and Executive Control, A. Miyake, Ed., pp. 62–101, Cambridge University Press, New York, NY, USA, 1999. View at: Google Scholar
  15. J. M. Chein, S. M. Ravizza, and J. A. Fiez, “Using neuroimaging to evaluate models of working memory and their implications for language processing,” Journal of Neurolinguistics, vol. 16, no. 4-5, pp. 315–339, 2003. View at: Publisher Site | Google Scholar
  16. D. LaBerge, “Sustained attention and apical dendrite activity in recurrent circuits,” Brain Research Reviews, vol. 50, no. 1, pp. 86–99, 2005. View at: Publisher Site | Google Scholar
  17. N. Cowan, D. Li, A. Moffitt et al., “A neural region of abstract working memory,” Journal of Cognitive Neuroscience, vol. 23, no. 10, pp. 2852–2863, 2011. View at: Publisher Site | Google Scholar
  18. R. Cabeza, “Role of parietal regions in episodic memory retrieval: the dual attentional processes hypothesis,” Neuropsychologia, vol. 46, no. 7, pp. 1813–1827, 2008. View at: Publisher Site | Google Scholar
  19. E. Ciaramelli, C. L. Grady, and M. Moscovitch, “Top-down and bottom-up attention to memory: a hypothesis (AtoM) on the role of the posterior parietal cortex in memory retrieval,” Neuropsychologia, vol. 46, no. 7, pp. 1828–1851, 2008. View at: Publisher Site | Google Scholar
  20. E. Ciaramelli, C. Grady, B. Levine, J. Ween, and M. Moscovitch, “Top-down and bottom-up attention to memory are dissociated in posterior parietal cortex: neuroimaging and neuropsychological evidence,” Journal of Neuroscience, vol. 30, no. 14, pp. 4943–4956, 2010. View at: Publisher Site | Google Scholar
  21. M. Wimber, H. J. Heinze, and A. Richardson-Klavehn, “Distinct frontoparietal networks set the stage for later perceptual identification priming and episodic recognition memory,” Journal of Neuroscience, vol. 30, no. 40, pp. 13272–13280, 2010. View at: Publisher Site | Google Scholar
  22. M. R. Uncapher and A. D. Wagner, “Posterior parietal cortex and episodic encoding: insights from fMRI subsequent memory effects and dual-attention theory,” Neurobiology of Learning and Memory, vol. 91, no. 2, pp. 139–154, 2009. View at: Publisher Site | Google Scholar
  23. A. D. Wagner, B. J. Shannon, I. Kahn, and R. L. Buckner, “Parietal lobe contributions to episodic memory retrieval,” Trends in Cognitive Sciences, vol. 9, no. 9, pp. 445–453, 2005. View at: Publisher Site | Google Scholar
  24. K. L. Vilberg and M. D. Rugg, “Memory retrieval and the parietal cortex: a review of evidence from a dual-process perspective,” Neuropsychologia, vol. 46, no. 7, pp. 1787–1799, 2008. View at: Publisher Site | Google Scholar
  25. R. Cabeza and L. Nyberg, “Imaging cognition II: an empirical review of 275 PET and fMRI studies,” Journal of Cognitive Neuroscience, vol. 12, no. 1, pp. 1–47, 2000. View at: Google Scholar
  26. K. A. Paller and A. D. Wagner, “Observing the transformation of experience into memory,” Trends in Cognitive Sciences, vol. 6, no. 2, pp. 93–102, 2002. View at: Publisher Site | Google Scholar
  27. D. Gaffan, “Widespread cortical networks underlie memory and attention,” Science, vol. 309, no. 5744, pp. 2172–2173, 2005. View at: Publisher Site | Google Scholar
  28. R. Henson, “A mini-review of fMRI studies of human medial temporal lobe activity associated with recognition memory,” Quarterly Journal of Experimental Psychology B, vol. 58, no. 3-4, pp. 340–360, 2005. View at: Publisher Site | Google Scholar
  29. E. I. Skinner and M. A. Fernandes, “Neural correlates of recollection and familiarity: a review of neuroimaging and patient data,” Neuropsychologia, vol. 45, no. 10, pp. 2163–2179, 2007. View at: Publisher Site | Google Scholar
  30. E. I. Skinner, M. A. Fernandes, and C. L. Grady, “Memory networks supporting effort retrieval success retrieval under conditions of full and divided attention,” Experimental Psychology, vol. 56, no. 6, pp. 386–396, 2009. View at: Publisher Site | Google Scholar
  31. M. E. Wheeler and R. L. Buckner, “Functional-anatomic correlates of remembering and knowing,” NeuroImage, vol. 21, no. 4, pp. 1337–1349, 2004. View at: Publisher Site | Google Scholar
  32. B. A. Shaywitz, S. E. Shaywitz, K. R. Pugh et al., “The functional neural architecture of components of attention in language-processing tasks,” NeuroImage, vol. 13, no. 4, pp. 601–612, 2001. View at: Publisher Site | Google Scholar
  33. A. L. W. Bokde, M. A. Tagamets, R. B. Friedman, and B. Horwitz, “Functional interactions of the inferior frontal cortex during the processing of words and word-like stimuli,” Neuron, vol. 30, no. 2, pp. 609–617, 2001. View at: Publisher Site | Google Scholar
  34. C. J. Fiebach, J. Rissman, and M. D'Esposito, “Modulation of inferotemporal cortex activation during verbal working memory maintenance,” Neuron, vol. 51, no. 2, pp. 251–261, 2006. View at: Publisher Site | Google Scholar
  35. J. A. Fiez and S. E. Petersen, “Neuroimaging studies of word reading,” Proceedings of the National Academy of Sciences of the United States of America, vol. 95, no. 3, pp. 914–921, 1998. View at: Publisher Site | Google Scholar
  36. A. N. Herbster, M. A. Mintun, R. D. Nebes, and J. T. Becker, “Regional cerebral blood flow during word and nonword reading,” Human Brain Mapping, vol. 5, no. 2, pp. 84–92, 1997. View at: Publisher Site | Google Scholar
  37. K. B. McDermott, S. E. Petersen, J. M. Watson, and J. G. Ojemann, “A procedure for identifying regions preferentially activated by attention to semantic and phonological relations using functional magnetic resonance imaging,” Neuropsychologia, vol. 41, no. 3, pp. 293–303, 2003. View at: Publisher Site | Google Scholar
  38. K. Hugdahl, R. Westerhausen, K. Alho, S. Medvedev, M. Laine, and H. HÄmÄlÄinen, “Attention and cognitive control: unfolding the dichotic listening story: cognition and neurosciences,” Scandinavian Journal of Psychology, vol. 50, no. 1, pp. 11–22, 2009. View at: Publisher Site | Google Scholar
  39. T. A. Christensen, S. M. Antonucci, J. L. Lockwood, M. Kittleson, and E. Plante, “Cortical and subcortical contributions to the attentive processing of speech,” Neuroreport, vol. 19, no. 11, pp. 1101–1105, 2008. View at: Publisher Site | Google Scholar
  40. G. H. Glover and C. S. Law, “Spiral-in/out BOLD fMRI for increased SNR and reduced susceptibility artifacts,” Magnetic Resonance in Medicine, vol. 46, no. 3, pp. 515–522, 2001. View at: Publisher Site | Google Scholar
  41. J. Talairach and P. Tournoux, Co-Planar Stereotaxic Atlas of the Human Brain, Thieme, New York, NY, USA, 1988.
  42. R. W. Cox, “AFNI: software for analysis and visualization of functional magnetic resonance neuroimages,” Computers and Biomedical Research, vol. 29, no. 3, pp. 162–173, 1996. View at: Publisher Site | Google Scholar
  43. A. Baddeley, “The fractionation of working memory,” Proceedings of the National Academy of Sciences of the United States of America, vol. 93, no. 24, pp. 13468–13472, 1996. View at: Publisher Site | Google Scholar
  44. C. Jacquemot and S. K. Scott, “What is the relationship between phonological short-term memory and speech processing?” Trends in Cognitive Sciences, vol. 10, no. 11, pp. 480–486, 2006. View at: Publisher Site | Google Scholar
  45. G. Hickok and D. Poeppel, “Dorsal and ventral streams: a framework for understanding aspects of the functional anatomy of language,” Cognition, vol. 92, no. 1-2, pp. 67–99, 2004. View at: Publisher Site | Google Scholar
  46. G. Hickok and D. Poeppel, “The cortical organization of speech processing,” Nature Reviews Neuroscience, vol. 8, no. 5, pp. 393–402, 2007. View at: Publisher Site | Google Scholar
  47. P. Belin, R. J. Zatorre, P. Lafaille, P. Ahad, and B. Pike, “Voice-selective areas in human auditory cortex,” Nature, vol. 403, no. 6767, pp. 309–312, 2000. View at: Publisher Site | Google Scholar
  48. J. Crinion, R. Turner, A. Grogan et al., “Language control in the bilingual brain,” Science, vol. 312, no. 5779, pp. 1537–1540, 2006. View at: Publisher Site | Google Scholar
  49. S. K. Scott, C. C. Blank, S. Rosen, and R. J. S. Wise, “Identification of a pathway for intelligible speech in the left temporal lobe,” Brain, vol. 123, no. 12, pp. 2400–2406, 2000. View at: Google Scholar
  50. S. K. Scott, “Auditory processing—speech, space and auditory objects,” Current Opinion in Neurobiology, vol. 15, no. 2, pp. 197–201, 2005. View at: Publisher Site | Google Scholar
  51. A. A. Zekveld, D. J. Heslenfeld, J. M. Festen, and R. Schoonhoven, “Top-down and bottom-up processes in speech comprehension,” NeuroImage, vol. 32, no. 4, pp. 1826–1836, 2006. View at: Publisher Site | Google Scholar
  52. M. D. Rugg, R. N. A. Henson, and W. G. K. Robb, “Neural correlates of retrieval processing in the prefrontal cortex during recognition and exclusion tasks,” Neuropsychologia, vol. 41, no. 1, pp. 40–52, 2003. View at: Publisher Site | Google Scholar
  53. P. C. Fletcher, T. Shallice, C. D. Frith, R. S. J. Frackowiak, and R. J. Dolan, “The functional roles of prefrontal cortex in episodic memory: II. Retrieval,” Brain, vol. 121, no. 7, pp. 1249–1256, 1998. View at: Publisher Site | Google Scholar
  54. K. B. McDermott, T. C. Jones, S. E. Petersen, S. K. Lageman, and H. L. Roediger, “Retrieval success is accompanied by enhanced activation in anterior prefrontal cortex during recognition memory: an event-related fMRI study,” Journal of Cognitive Neuroscience, vol. 12, no. 6, pp. 965–976, 2000. View at: Publisher Site | Google Scholar
  55. M. L. Seghier, E. Fagan, and C. J. Price, “Functional subdivisions in the left angular gyrus where the semantic system meets and diverges from the default network,” Journal of Neuroscience, vol. 30, no. 50, pp. 16809–16817, 2010. View at: Publisher Site | Google Scholar
  56. J. T. Becker, D. K. MacAndrew, and J. A. Fiez, “A comment on the functional localization of the phonological storage subsystem of working memory,” Brain and Cognition, vol. 41, no. 1, pp. 27–38, 1999. View at: Publisher Site | Google Scholar
  57. S. M. Ravizza, M. R. Delgado, J. M. Chein, J. T. Becker, and J. A. Fiez, “Functional dissociations within the inferior parietal cortex in verbal working memory,” NeuroImage, vol. 22, no. 2, pp. 562–573, 2004. View at: Publisher Site | Google Scholar
  58. V. J. Schmithorst, S. K. Holland, and E. Plante, “Diffusion tensor imaging reveals white matter microstructure correlations with auditory processing ability,” Ear and Hearing, vol. 32, pp. 156–167, 2011. View at: Publisher Site | Google Scholar
  59. M. E. Wheeler and R. L. Buckner, “Functional dissociation among components of remembering: control, perceived oldness, and content,” Journal of Neuroscience, vol. 23, no. 9, pp. 3869–3880, 2003. View at: Google Scholar
  60. J. A. Fiez, E. A. Raife, D. A. Balota, J. P. Schwarz, M. E. Raichle, and S. E. Petersen, “A positron emission tomography study of the short-term maintenance of verbal information,” Journal of Neuroscience, vol. 16, no. 2, pp. 808–822, 1996. View at: Google Scholar
  61. T. Shallice, P. Fletcher, C. D. Frith, P. Grasby, R. S. J. Frackowiak, and R. J. Dolan, “Brain regions associated with acquisition and retrieval of verbal episodic memory,” Nature, vol. 368, no. 6472, pp. 633–635, 1994. View at: Publisher Site | Google Scholar
  62. J. Jonides, E. H. Schumacher, E. E. Smith et al., “The role of parietal cortex in verbal working memory,” Journal of Neuroscience, vol. 18, no. 13, pp. 5026–5034, 1998. View at: Google Scholar
  63. R. J. Zatorre, M. Bouffard, P. Ahad, and P. Belin, “Where is “where” in the human auditory cortex?” Nature Neuroscience, vol. 5, no. 9, pp. 905–909, 2002. View at: Publisher Site | Google Scholar
  64. J. D. Johnson, S. G. R. McDuff, M. D. Rugg, and K. A. Norman, “Recollection, familiarity, and cortical reinstatement: a multivoxel pattern analysis,” Neuron, vol. 63, no. 5, pp. 697–708, 2009. View at: Publisher Site | Google Scholar
  65. N. Cowan and N. L. Wood, “Constraints on awareness, attention, processing, and memory: some recent investigations with ignored speech,” Consciousness and Cognition, vol. 6, no. 2-3, pp. 182–203, 1997. View at: Publisher Site | Google Scholar
  66. N. Cowan, “Activation, attention, and short-term memory,” Memory and Cognition, vol. 21, no. 2, pp. 162–167, 1993. View at: Google Scholar

Copyright © 2012 Thomas A. Christensen et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.