OUP user menu

The Enigmatic temporal pole: a review of findings on social and emotional processing

Ingrid R. Olson, Alan Plotzker, Youssef Ezzyat
DOI: http://dx.doi.org/10.1093/brain/awm052 1718-1731 First published online: 28 March 2007


The function of the anterior-most portion of the temporal lobes, the temporal pole, is not well understood. Anatomists have long considered it part of an extended limbic system based on its location posterior to the orbital frontal cortex and lateral to the amygdala, along with its tight connectivity to limbic and paralimbic regions. Here we review the literature in both non-human primates and humans to assess the temporal pole's putative role in social and emotional processing. Reviewed findings indicate that it has some role in both social and emotional processes, including face recognition and theory of mind, that goes beyond semantic memory. We propose that the temporal pole binds complex, highly processed perceptual inputs to visceral emotional responses. Because perceptual inputs remain segregated into dorsal (auditory), medial (olfactory) and ventral (visual) steams, the integration of emotion with perception is channel specific.

  • perirhinal cortex
  • anterior temporal lobe
  • ba 38
  • face processing
  • frontotemporal dementia


There has been considerable attention devoted to understanding the neural basis of social and emotional processing. The appeal of understanding these processes is driven partially by a need to understand disrupted socioemotional processing that lies at the heart of many psychiatric disorders. The two neuroanatomical regions most commonly linked to socioemotional processing are the amygdala and orbital portions of the prefrontal cortex. Less attention has been paid to a third region that lies between the orbital frontal cortex and the amygdala and receives and sends connections to both regions: the temporal pole (TP). Anatomists have referred to the TP as a paralimbic region based on its anatomical location and connectivity (Duvernoy, 1999; Mesulam, 2000). Whether the human TP has any social or emotional functions is not well-understood. An influential review of the fMRI literature noted that little was known about the functionality of this region and that few fMRI studies had reported activation there (Cabeza and Nyberg, 2000). This lack of knowledge is surprising given the fact that for decades, neurosurgeons have removed this area, along with diseased medial temporal lobe regions, in epilepsy resection surgery. Here we review the literature on the TP in hopes of filling an anatomical void in the literature on socioemotional functions and the brain. We begin by discussing TP anatomy, and then move on to discuss Klüver–Bucy syndrome in monkeys and humans, disordered social and emotional processing in humans, face processing and theory of mind. We end with a theoretical discussion that synthesizes evidence on the particular role of the TP is socioemotional processing.

Temporal pole anatomy

There are several excellent reviews of the anatomy of the anterior temporal lobe (Nakamura and Kubota, 1996; Stefanacci et al., 1996; Gloor, 1997; Mesulam, 2000; Kondo et al., 2003, 2005). that we summarize here. The temporal pole, also referred to as BA 38, planum polare, area TG or the anterior aspect of perirhinal cortex, covers the anterior-most end of the temporal lobe, somewhat like a cap (Fig. 1). It is rostral to the perirhinal cortex. Some investigators have included the TP within their definition of the perirhinal cortex (Insausti et al., 1987a, b; Suzuki and Amaral, 1994), which traditionally encompasses areas 35 and 36; however, it is more commonly considered a separate region (Insausti et al., 1998).

Fig. 1

(A) Brodmann's depiction of the lateral surface of the human brain. The temporal pole (BA 38) is highlighted in orange. (B) A schematic diagram of the inferior surface of the human brain showing the anatomical relationship of the temporal pole to the posterior orbital frontal cortex and the amygdala (in reality the amygdala is buried under the cortex, but for the sake of illustration, has been depicted on the surface).

The TP is highly interconnected with both the amygdala and orbital frontal cortex and is therefore often referred to as a paralimbic region. A large white matter tract, the uncinate fascicule, links the TP to prefrontal regions. Like other limbic regions, the TP receives and sends projections to the basal forebrain. It has been noted that the pattern of connectivity of the TP bears striking similarity to that of the amygdala. Dorsal portions of the TP project to the hypothalamus, a neuromodulatory region important for autonomic regulation of emotions. The tight connection between the TP and the hypothalamus may explain why electrical stimulation of the human TP produces changes in heart rate, respiration and blood pressure (Gloor et al., 1982).

The TP, along with the temporal–parietal junction, has also been described as association cortex due to its unusual connectivity. It sends and receives connections to the three sensory systems represented in the temporal lobe. In the macaque, the dorsolateral TP receives projections from third-order auditory association cortex, the ventral TP receives projections from extrastriate visual cortex in the inferiortemporal lobe, and the medial TP receives projections from prepiriform olfactory cortex in medial aspects of the temporal lobe. Furthermore, the medial TP receives projections from the insula, which has an important role in gustation and awareness of internal physiological state (Critchley, 2004). Existing evidence from the macaque indicates that these processing streams do not converge, but rather, remain separate in the TP.

Because of a lack of anatomical data, it is unclear how the connectivity of the human TP compares to that of the macaque (Nakamura and Kubota, 1996). The language functions that are present in human temporal lobe, for example, make it possible that the connectivity of the human TP differs significantly from that found in non-human primates. As a first step in investigating this possibility, we reviewed neuroimaging studies that reported TP activations to a wide variety of tasks and stimuli and plotted the coordinates of anterior temporal lobe activations on a standard brain. We found that these activations tended to follow a dorsal/ventral segregation based upon whether the stimuli used were auditory or visual (Fig. 2A; methods found in Appendix A). This finding provides one piece of evidence that the human TP has a similar pattern of connectivity as the non-human primate TP. These findings also suggest that one function of the human TP is multimodal perceptual analysis. Its anatomical location and connectivity suggest that its function extends beyond perception to social and emotional processing, reviewed next.

Fig. 2

Centres of activation from fMRI and PET reporting anterior temporal lobe activations to (A) auditory or visual stimuli. Red dots show centres of activation reported for tasks that used clearly visual stimuli such as faces or houses, whereas blue dots show centres of activation reported for tasks that used auditory stimuli such as voices or music. (B) Emotion or theory of mind tasks. Red dots show centres of activation reported in emotion tasks, whereas blue dots show centres of activation reported in theory of mind tasks. Activation maps were created with MRIcro. Activation foci are shown in MNI space, in radiological convention (e.g. left is on the right). Studies used to create these images are listed in Appendices A and B.

Temporal pole and Klüver–Bucy syndrome

One of the first hints that the TP was involved in socioemotional processing came from studies of Klüver–Bucy syndrome. The Klüver–Bucy syndrome in monkeys consists of tameness and diminished fear, hyperorality, hypersexuality, blunted affect, visual agnosia and social withdrawal (Kluver and Bucy, 1939; Kling et al., 1993). Later studies found that lesions to the monkey TP, orbital frontal cortex or amygdala produce most, but not all, of the Klüver–Bucy symptoms (Horel et al., 1975; Kling and Steklis, 1976; Kling et al., 1993). The exception to this is ‘psychic blindness’ or visual agnosia, which is only produced when lesions include more posterior portions of the temporal lobe. Lesions to nearby areas, including inferior temporal cortex, superior temporal cortex, lateral surface of the frontal lobe and the anterior cingulate do not produce the striking social and emotional deficits of the Klüver–Bucy syndrome (Kling and Steklis, 1976; Kling et al., 1993).

Perhaps the most interesting symptom of the Klüver–Bucy disorder is ‘social withdrawal’, a description that minimizes the scope of social impairment. Female monkeys with surgical lesions of the TP, excluding the amygdala, exhibit grossly abnormal social behaviour (Bucher et al., 1970; Myers and Swett, 1970; Franzen and Myers, 1973; Myers, 1975; Kling and Steklis, 1976; Kling et al., 1993). They do not produce appropriate social signals (vocal or facial) nor do they appear to recognize the social signals of peer monkeys. They showed little social interest in their peers, and at times are rejected from their social group. They become tame and show little aggression towards peer monkeys when provoked. Those with babies are neglectful and often violent towards them, causing consternation among other female monkeys. Some, but not all, of these social problems improve over time. Prefrontal and amygdala lesions lead to a similar pattern of behaviour, underscoring the tight coupling of the TP to these regions. In contrast, lesions to posterior temporal cortex (TE or inferior temporal cortex) cause no social problems. (Table 1). These findings have been replicated in other species of monkey (Kling and Steklis, 1976), suggesting that there is evolutionary conservation of structure-function in the TP across primate species.

View this table:
Table 1

Summary of social and emotional behavioural change following bilateral ablation of orbital portions of the prefrontal cortex (PFC), temporal pole (TP), amygdala, inferior temporal cortex (IT) or anterior cingulated cortex (ACC) in various species of adult Old World monkeys

    Diminished fearxxx
    Blunted affectxxxx
    Visual Agnosiax
    Social disinterestxxxx
Decreased production of facial and vocal signalsxxx
Decreased aggressionxxx
Rejection from social groupxxx
Decreased social statusxxx
Aberrant maternal behaviourxxx
Decreased groomingxxx
  • Note: It should be noted that Klüver–Bucy symptoms were only observed in laboratory monkeys while free-ranging monkeys tended to instead show severe social disinterest and withdrawal. Also, ablations in infant monkeys tend to cause less marked social deficits. Evidence derived from (Bucher et al., 1970; Franzen and Myers, 1973; Kling and Steklis, 1976; Kling et al., 1993; Miller et al., 1995, 1997; Mychack et al., 2001; Amaral et al., 2003; Hadland et al., 2003; Thompson et al., 2003).

One explanation that has been offered for these findings is that prefrontal and TP regions are instrumental in supporting group social behaviours such as the tendency to band together to form societies, to participate in family life and to interact with others in the social group (Myers, 1969). However, there are many potential mental processes that sustain group social behaviours and no effort has been made to disentangle which of them are disrupted with TP damage. For instance, problems in recognizing the meaning of social signals, in associating a social signal with the proper response or with an emotional reaction, in social interest, or in feeling emotions could all cause social deviance. Nevertheless, these findings constitute the first piece of evidence that the anterior temporal lobes, including the TP, play some important role in socioemotional processing.

Temporal pole and clinical disorders of socioemotional regulation

TP dysfunction in humans has been associated with a host of socioemotional disorders. Klüver–Bucy syndrome can be observed after bilateral medial and anterior temporal lobe damage, as often occurs in herpes encephalitis (Lilly et al., 1983). It has been observed after unilateral anterior temporal lobe damage (Ghika-Schmid et al., 1995) and as a post-ictal temporal lobe epilepsy phenomenon (Anson and Kuhlman, 1993). It can also be observed in dementing diseases that affect the anterior-medial temporal lobes (Lilly et al., 1983).

One of the most interesting deficits in socioemotional processing arises after atrophy to the right anterior temporal lobe, typically seen in the temporal variant of frontal temporal dementia (tv-FTD). FTD is a progressive disease characterized by a somewhat rapid degeneration of frontal and/or anterior temporal lobe tissue that can be left or right lateralized, and frontal or temporal localized. Temporal damage is most evident in polar regions, while more medial regions, such as the hippocampus, remain intact at early stages of the disease (Mummery et al., 2000; Hodges, 2001). Patients with right, but not left, TP atrophy due to the tv-FTD exhibit changes in personality and socially appropriate behaviour (Thompson et al., 2003). For instance, Gorno-Tempini and colleagues (2004) studied the cognitive, behavioural and anatomical features of patient JT who presented with marked changes in behaviour and personality. JT changed from an extraverted, highly empathic individual to a somewhat introverted and cold individual, lacking in empathy. She also lost much of her social dominance and became neurotic and demanding. Her indiscriminate eating behaviours provided the clearest evidence of her socially inappropriate behaviour. ‘If not controlled, [she] would eat floral table decorations and large quantities of butter, oil and jam by themselves’ (Gorno-Tempini et al., 2004).

As this example illustrates, tv-FTD can at times cause partial or full-blown Klüver–Bucy syndrome. In fact, many of the deficits found in right lateralized tv-FTD mimic deficits observed in monkeys with surgical ablation of bilateral TP (Table 2). It has been reported that these patients have fixed facial expressions and have difficulties in posing facial expressions. Their abnormal expressions of affect make social interactions uncomfortable for both strangers and family members. Other affect-related problems include depression, irritability, apathy, emotional blunting and being ill at ease in company (Miller et al., 1995; Miller et al., 1997; Bozeat et al., 2000; Mychack et al., 2001; Thompson et al., 2003). It is likely that some of the Klüver–Bucy symptoms resulting from bilateral TP damage are due to severing of white matter connections to other brain regions. Enhancement of oral tendencies has been produced experimentally by ablation of orbital frontal cortex (Butter et al., 1969), and combined hyperphagia and hypersexuality has been observed in humans with thalamic lesions (Poeck and Pilleri, 1965). Both regions are highly interconnected with the TP (Mesulam, 2000).

View this table:
Table 2

A list of social and emotional deficits exhibited by monkeys with surgical ablation of bilateral TP, sparing the amygdala and analogous social and emotional deficits observed in humans with right lateralized tv-FTD

Monkeys with TP ablationHumans with right tv-FTD
Loss of production of facial or vocal signallingLoss of production of facial expressions; monotone voice
Loss of recognition of peer facial or vocal signalsLoss or poor recognition of facial expressions
Decreased interest in peersSelf-centred; loss of empathy; preoccupation with non-social stimuli (e.g. jigsaw puzzles)
Decreased aggressionDecreased affect
Rejection from social groupAvoided by family and friends
Decreased status in social hierarchyLoss of extraversion and social dominance
Aberrant and neglectful maternal behaviour
Aberrant sexual behaviourHypo or hyper-sexuality
Decreased groomingDecreased grooming; bizarre changes in dress
  • Note: Evidence derived from (Bucher et al., 1970; Franzen and Myers, 1973; Kling and Steklis, 1976; Kling et al., 1993; Miller et al., 1995, 1997; Mychack et al., 2001; Thompson et al., 2003).

TP damage can lead to unstable mood states. Damage due to trauma or surgical resection can lead to rapidly cycling bipolar disorder (Murai and Fujimoto, 2003) or rapidly cycling changes in clinical levels of depression, anxiety and irritation (Glosser et al., 2000). The link between vascilating mood states and TP damage is also found in reports that traumatic or surgical damage to the anterior temporal lobe damage is associated with acquired bipolar disorder (Jorge et al., 1993; Brooks and Hoblyn, 2005) and mania (Carran et al., 2003). Last, there are intriguing links between anterior temporal lobe dysfunction and schizophrenia: several studies have reported that schizophrenics have smaller than normal temporal poles (Gur et al., 2000; Kasai et al., 2003; Crespo-Facorro et al., 2004) and that TP neurons in post-mortem schizophrenic brain samples have abnormal microstructure (Ong and Garey, 1993). Schizophrenia-like psychoses, when present in epileptics, are most commonly observed when the seizure foci is in the left anterior temporal lobe (Glosser et al., 2000).

Temporal pole and face processing

The ability to recognize and identify other people is fundamental to human social processing. Numerous studies have identified a region of temporal cortex near the superior temporal sulcus in non-human primates that contains cells responsive to faces (Perrett et al., 1985; Desimone, 1991; Gross, 1992) and an analogous region has been identified in inferior portions of the human temporal lobe (Kanwisher et al., 1997). Fewer investigators are aware of evidence implicating the TP in face processing. There are four sources of evidence.

First, bilateral lesions to the anterior temporal lobes can cause a particular type of prosopagnosia termed ‘amnesic associative prosopagnosia’ by Damasio and colleagues (1990; see Fig. 3). Its name is based on the observation that it often occurs in the context of amnesia (Damasio et al., 1990). Two distinguishing features of this type of prosopagnosia are that patients’ have preserved perception of faces, but impaired recognition of faces, and their deficit extends to non-face recognition cues like voice and gait. Bilateral damage and memory loss may not be necessary conditions for this disorder, as similar profiles of spared face perception and impaired face recognition have been reported for cases in which memory is intact and anterior temporal lobe damage is unilateral (Sergent and Signoret, 1992; Barton et al., 2001, 2004).

Fig. 3

A schematic diagram of distribution of damage found in cases of amnesic associative prosopagnosia, also termed associative agnosia. The lesions shown in this diagram are bilateral and involve the anterior but not posterior temporal lobe (reprinted from Damasio et al. 1990, with permission).

Anterior temporal lobe resection surgery can also produce face processing deficits. Findings from this line of research have revealed interesting laterality differences. Left resection surgery can cause impaired proper naming abilities when shown photos of famous faces, while right resection surgery can cause impaired recognition of famous (Glosser et al., 2003) or personally familiar (Tippett et al., 2000) faces. Such deficits are observed whether naming people from a picture or a verbal description (Fukatsu et al., 1999; Glosser et al., 2003; Tsukiura et al., 2003). Left TP resection can impair the ability to learn new face–name associations, whereas this is not true of right TP resections (Tsukiura et al., 2003).

Second, temporal pole atrophy, as seen in tv-FTD, can cause severe impairments in naming famous faces. Thompson et al. (2003) found that 31 of the 47 FTD patients in their study had specific complaints of difficulty recalling people's names. As with temporal lobe resection cases, name recognition difficulties are more commonly associated with left lateralized damage, while face recognition difficulties are more common after right-lateralized damage (Snowden et al., 2004). This type of face processing deficit has been termed ‘progressive prosopagnosia’ (Barbarotto et al., 1995; Evans et al., 1995; Gentileschi et al., 1999, 2001; Gainotti et al., 2003; Joubert et al., 2003; Snowden et al., 2004).

Third, two studies in which face-related event-related potentials (ERPs) were recorded from electrodes implanted in the human brain reported that late face-specific ERPs (around 350 ms) were found in the ventral TP. These ERPs were later than face ERPs found in more posterior visual areas, and were right-lateralized whereas fusiform ERPs were bilateral (Allison et al., 1994, 1999).

Fourth, a few neuroimaging studies have reported activations in the TP to faces. Some studies have reported increased activations to familiar faces as compared to unfamiliar faces (Gorno-Tempini et al., 1998; Leveroni et al., 2000; Nakamura et al., 2000; Grabowski et al., 2001; Pourtois et al., 2005; Rothstein et al., 2005; Sergent et al., 1992; Sugiura et al., 2001) which concurs with lesion studies showing that the anterior temporal lobe is critical for person, but not face, recognition (Damasio et al., 1990). Other studies have reported increased activations to emotional faces as compared to neutral faces (Phillips et al., 1998; Blair et al., 1999; Tsukiura et al., 2003; Kim et al., 2005).

In sum, evidence from numerous sources indicates that the anterior temporal lobe, though not critical for the perceptual analysis of faces, is critical for linking person-specific memories to perceptual representations of faces. Clearly, such a process is critical for adept social interactions and as such, may help to explain some of the deviant social behaviour exhibited by monkeys with TP lesions, reviewed earlier.

Is the TP selective for face processing? There are reports of patients with circumscribed TP damage that cannot recognize celebrities or name them from descriptions but have unimpaired recognition of landmarks (Gainotti et al., 2003; Damasio et al., 2004; Tranel, 2006). However, while face recognition difficulties are usually the earliest and most prominent symptom of TP damage due to progressive prosopagnosia (Barbarotto et al., 1995; Evans et al., 1995; Gentileschi et al., 2001; Gainotti et al., 2003; Thompson et al., 2003; Snowden et al., 2004), difficulties often extend to other modalities, such as recognizing people from their voice (Gentileschi et al., 2001; Gainotti et al., 2003), their name (Evans et al., 1995; Snowden et al., 2004) and even by their handwriting (Gentileschi et al., 2001). In other words, unlike prosopagnosics with more posterior damage, progressive prosopagnosics lose all access to stored representations of person-related knowledge, a type of semantic memory.

Progressive prosopagnosics can exhibit other types of recognition problems too, such as an inability to recognize famous monuments and songs (Barbarotto et al., 1995; Gentileschi et al., 2001). Neuroimaging studies have largely confirmed these findings: two PET studies reported overlapping right temporal pole activations in response to viewing familiar faces and familiar buildings (Nakamura et al., 2000; Grabowski et al., 2001).

These findings can be reconciled if one takes into account the severity of impairment. It has been noted that TP patients with the worst person recognition scores also have difficulty with other non-person proper names (Hanley and Kay, 1998). A similar explanation can be applied to tv-FTD patients with person-identification deficits since patients with selective deficits identifying people (Evans et al., 1995; Gainotti et al., 2003) are nevertheless, quantitatively better at face-recognition than are patients whose impairments extend to other unique items (Barbarotto et al., 1995; Gentileschi et al., 2001). This finding suggests that the TP is necessary, but not specialized, for person recognition. Person recognition may simply be more sensitive to neuronal loss in the TP than are other types of subordinate-level recognition.

Thinking about others: theory of mind

Theory of mind is the ability to infer the desires, intentions or beliefs of others. There are several ways to study this, although one of the most common ways is to read short stories or view cartoons that depict a scenerio in which another's thoughts or beliefs must be inferred. Many neuroimaging studies (Table 3), using different tasks and stimuli, have implicated the temporal poles in theory of mind (see Fig. 2B for activations patterns) (Fletcher et al., 1995; Goel et al., 1995; Baron-Cohen et al., 1999; Brunet et al., 2000; Gallagher et al., 2000; McCabe et al., 2001; Vogeley et al., 2001; Berthoz et al., 2002; Ferstl and von Cramon, 2002; Gallagher et al., 2002; Calarge et al., 2003; Walter et al., 2004; Mitchell et al., 2006; Saxe and Powell, 2006). An implicit theory of mind task in which subjects viewed geometric shapes that moved with intentionality and cause (Heider and Simmel, 1944) has been used by several groups who also found activations in the temporal pole (Castelli et al., 2000; Schultz et al., 2003; German et al., 2004; Iacoboni et al., 2004; Ohnishi et al., 2004).

View this table:
Table 3

A list of neuroimaging studies of mentalizing or theory of mind

StimuliTP activation?
Berthoz (2002)
Blakemore (2003)
Brunet (2000)
Calarge (2003)
Calder (2002)
Castelli (2000)
Ferstl (2002)
Fletcher (1995)
Gallagher (2000)
Gallagher (2002)
German (2004)
Grezes (2004)
Iacoboni (2004)
Mitchell (2006)
Moll (2002)
Ohnishi (2004)
Rilling (2004)
Saxe (2006)
Schultz (2003)
Vogley (2001)
Walter (2004)
  • Note: Studies that used head coils are excluded from the list. The type of visual stimulus is listed in the middle columns, and the far right columns indicate whether or not temporal pole activations were observed.

Other tasks that require subjects to think about others’ thoughts and emotions have produced TP activations. One study reported TP activations when subjects were asked to detect deception (Grezes et al., 2004), while another study reported TP activations when subjects were asked to make moral decisions (Moll et al., 2002b; Heekeren et al., 2003). At least three studies have reported activations in the temporal pole while inferring the emotional state of others (Farrow et al., 2001; Carr et al., 2003; Vollm et al., 2006). TP activations correlate with ‘personal distress scores’, a measure of how much you feel another's negative emotions, in normal populations of college students (Moriguchi et al., 2006). Interestingly, Vollm and colleagues (2006) found that activation clusters for a theory of mind task and an empathy task overlapped in the TP, leading them to suggest that this region functions generally in making inferences about the mental state of others. These findings have led some researchers to conclude that the temporal poles are a critical part of the ‘mentalizing’ brain (Frith, 2001; Frith and Frith, 2003).

Neuropsychological studies have provided mixed results for this theory. One of the only neuropsychological studies that examined this topic tested a large sample of patients with various degrees of amygdala damage. The results showed that adult-onset amygdala damage did not generally impair theory of mind performance and importantly, that extra-amygdala damage to the TP did not compromise theory of mind abilities (Shaw et al., 2004). This finding casts some doubt on the idea that the TP (or the amygdala) is critical for theory of mind in adults, although it should be noted that the TP was only partially damaged in all cases. In contrast, evidence from tv-FTD lends support to the theory that the TP is critical for theory of mind. A common complaint among family members of patients with tv-FTD is that the patient exhibits a pervasive lack of empathy (Mychack et al., 2001). Recent findings show that patients with FTD have lower levels of empathy than controls or patients with Alzheimer's disease (Rankin et al., 2006) and that tv-FTD is associated with disruption of both emotional empathy and perspective-taking, one measure of cognitive empathy that is similar to theory of mind (Rankin et al., 2006).

Various ideas have been offered for how to interpret these findings. One suggestion is that TP activations in theory of mind tasks reflect retrieval of semantic memory scripts (Frith and Frith, 2003); however, there is no reason to think that the semantic memory demands of theory of mind tasks are any different from the semantic memory demands of non-theory of mind control tasks. This is especially true of theory of mind tasks that used geometric shapes that move with intentionality and cause (Castelli et al., 2000; Schultz et al., 2003). Another interpretation is that the TP has some role in encoding personal memories (Nakamura and Kubota, 1995) and thus in theory of mind tasks, the TP utilizes personal memories to comprehend the state of mind of others (Moriguchi et al., 2006). Last, it is possible that theory of mind tasks taps person recognition functions of the TP, although this idea does not account for the fact that most studies of face processing and gaze processing have failed to report activations in the TP. Further consideration of this topic is found in the section titled ‘Negative Findings’.

The temporal pole in socioemotional processing: theoretical considerations

The studies reviewed in the previous sections provide convincing evidence that the TP plays some role in socioemotional processing. We now turn to the problem of defining what exactly that role is. We have already provided some analysis of the function of this region for face processing and theory of mind processing; here we offer a more general framework for understanding the social and emotional functionality of the TP. It has been suggested that the TP is part of a system that modulates visceral emotional functions in response to emotionally evocative perceptual stimuli, based on its anatomical connectivity (Kondo et al., 2003, 2005). We further suggest that this system is reactivated when emotions are perceived or imagined and that mnemonic functions of this region allow for storage of perception–emotion linkages, forming the basis of personal semantic memory.

The macaque dorsal TP receives input from auditory association areas (Kondo et al., 2003), cells in this region are known to respond to complex auditory stimuli (Kondo et al., 2003; Poremba et al., 2003), and as shown in Fig. 2A, auditory stimuli tend to activate dorsal portions of the anterior temporal lobe, with a left-lateralization bias. If our hypothesis is true, it is possible that dorsal portions of the TP are responsible for coupling visceral emotional responses with representations of complex auditory stimuli. A case study hints at this possibility. It was reported that in the year after having left temporal lobectomy surgery, a young man exhibited dramatic changes in musical taste. ‘He found that the [rock] music he used to listen to before the operation sounded ‘‘too hard, too fast, and too violent”. He now had a preference for Celtic or Corsican polyphonic singing’ and ‘…he now had difficulty staying with his old friends, since he could no longer share his musical preferences and hence his topics of discussion’ (Sellal et al., 2003). Thus loss of the left TP led to changes in emotional reactions to complex auditory stimuli and, subsequently, to changes in social affiliation. At least two neuroimaging studies have reported activations in this region to aversive sounds such as a baby crying (Lorberbaum et al., 2002) or woman screaming (Royet et al., 2000). Many neuroimaging studies have reported activations in this region to pleasant sounds such as complex music (Brown et al., 2004; Platel, 2005; Satoh et al., 2006) and laughter (Royet et al., 2000).

In contrast, the macaque ventral TP receives inputs from visual association areas in the inferior temporal lobe (Kondo et al., 2003) and is thought to be the endpoint of visual processing. Cells in the macaque TP respond most strongly to complex visual stimuli (Nakamura et al., 1994; Nakamura and Kubota, 1996) and as shown in Fig. 2A, ventral portions of the human TP tend to respond to complex visual stimuli such as faces, cartoons and photographs of houses in neuroimaging tasks. Our hypothesis would predict that ventral portions of the TP couple visceral emotional responses to complex visual stimuli. Sellal (2003) reported that the lobectomy patient who experienced a radical change in musical tastes, also exhibited changes in pictoral preferences, a finding that has also been noted in the FTD literature (Miller et al., 1998, 2000). Some PET and fMRI studies have used visual stimuli to evoke negative emotions and found activations in the TP to emotions such as sadness (Lane et al., 1997b; Blair et al., 1999; Eugene et al., 2003; Levesque et al., 2003), anxiety (Reiman et al., 1989; Chua et al., 1999; Kimbrell et al., 1999), anger (Dougherty et al., 1999; Kimbrell et al., 1999; Damasio et al., 2000), fear (Phillips et al., 1998) and disgust (Lane et al., 1997b; Phillips et al., 1998). TP activations have also been associated with various positive emotions such as humour generated by cartoons (Mobbs et al., 2003), sexual arousal induced by pornography (Redoute et al., 2000; Beauregard et al., 2001) or maternal feeling generated by watching a video of one's own infant (Ranote et al., 2004).

Medial aspects of the macaque TP receive olfactory and gustatory inputs (Mesulam, 2000). Few studies have examined emotional processing of these stimuli in humans; there is some evidence from PET that the TP is responsive to valenced olfactory stimuli (Royet et al., 2000).

One prediction of our hypothesis is that damage to the TP should cause a decoupling of high-level perception with visceral emotional experience. One study reported that a patient who underwent right temporal lobectomy lost all emotional attachments to his family members, although recognition of them remained intact (Lipson et al., 2003). It has also been reported that a patient with a large right TP lesion extending posteriorly in the fusiform gyrus believed that her family had been replaced by imposters (Hudson and Grace, 2000), possibly due to a decoupling between neural face recognition systems and emotion systems (Hirstein and Ramachandran, 1997). Last, as noted earlier, monkeys with bilateral TP ablations lose normal emotional attachments to their infants and to peer monkeys (Bucher et al., 1970; Kling and Steklis, 1976).

A question that arises is how does the TP differ functionally from the amygdala or the orbital frontal cortex? One strong possibility is that within the TP there is still sensory-limbic segregation of auditory, visual and olfactory channels that is not present in the amygdala or orbital frontal cortex (Kling and Steklis, 1976). This raises the possibility of modality-specific socioemotional disorders. Resection for temporal lobe epilepsy would cause unilateral damage to the entire TP. However, trauma, stroke or tumour could cause focal damage to only one sensory-limbic channel.

Negative findings

There is a growing neuroimaging literature on socioemotional processing. Most of these studies report amygdala and various frontal activations, but do not report TP activations. One difficulty in interpreting null results is that the lack of activations could either mean that the TP is insensitive to the task or stimulus or it could simply reflect the fact that air-tissue inhomogeneities, a problem in imaging all neural regions near the sinuses, have diminished the BOLD signal to such a degree that no activations are visible (Devlin et al., 2000). An additional problem is that the existing literature on socioemotional processing has at times emphasized activations in the amygdala and frontal regions but downplayed activations in regions such as the TP. TP activations are frequently listed in tables but are not discussed (Beauregard et al., 2001; Moll et al., 2002b), or are lumped into ill-defined anatomical categories such as ‘periamygdaloid’ or ‘paralimbic cortex’ (Miller et al., 2005; Dapretto et al., 2006), which tends to obscure the reader's ability to discern TP activations.

These issues aside, it is interesting to note that TP activations are frequently observed in complex emotional tasks such as theory of mind tasks, but less frequently observed in simpler emotional tasks, such as emotional face perception or gaze perception tasks. Table 3 lists neuroimaging studies of theory of mind and shows that the TP is activated by a variety of stimuli: movies, comics, sentences and stories. The instances in which it is not activated are informative. For instance no activations were observed in one study that used animated sequences that lacked a social component (Blakemore et al., 2003), whereas animated sequences that contain a social component are overwhelmingly associated with TP activations (Castelli et al., 2000; Schultz et al., 2003; German et al., 2004; Grezes et al., 2004; Iacoboni et al., 2004; Ohnishi et al., 2004). Although a social component appears to be necessary to activate this region, it is not sufficient. Two studies in which comparisons were made between conditions in which subjects thought they were playing computer games against humans versus computers did not report TP activations (Gallagher et al., 2002; Rilling et al., 2004) nor have studies that have examined various aspects of eye gaze (Calder et al., 2002).

The findings to date from neuroimaging studies of theory of mind suggest that the TP is sensitive to stimuli that tell a story. More specifically, the TP is sensitive to stimuli with socially important narratives, either in the form of a film strip, a comic strip or a story, and to tasks that require one to analyse other agent's emotions, intentions or beliefs. It is somewhat insensitive to stimuli with non-social narratives and to tasks that require simple perceptual-level analyses such as direction of gaze.

As mentioned earlier, bilateral removal of monkey TP disrupts the ability to recognize, interpret and react to a host of social signals (Kling and Steklis, 1976). This region contains cells that are responsive to complex visual and auditory stimuli (Nakamura and Kubota, 1996; Kondo et al., 2003, 2005). These cells are critical for both high-level perception and perceptual memory because they demonstrate certain mnemonic properties (Nakamura and Kubota, 1996). One can therefore interpret the neuroimaging findings from theory of mind tasks as reflecting the linkage of recognized social cues to emotional interpretations and reactions. Unlike regions such as the superior temporal sulcus, which is sensitive to biological motion and direction of gaze (Puce and Perrett, 2003), the TP is sensitive to complex social stimuli with a narrative or script.

Laterality differences

There is some evidence pointing towards interesting laterality differences in the TP. The right anterior temporal lobe appears to be associated with emotion and socially relevant memory. Right lateralized temporal lobe epilepsy is associated with a higher prevalence (83 versus 69%) of axis-I psychiatric disorders such as major depression and anxiety disorders as compared to left lateralized temporal lobe epilepsy (Glosser et al., 2000). Atrophy of the right anterior temporal lobe in tv-FTD is associated with changes in mood (e.g. emotional blunting, depression, irritability, apathy), personality (e.g. from extraverted to introverted; sudden acquisition of peculiar new interests and hobbies) and in socially appropriate behaviour (Miller et al., 1995, 1997; Bozeat et al., 2000; Mychack et al., 2001; Thompson et al., 2003). The degree to which these social problems reflect temporal or frontal damage is difficult to estimate since only macroscopic tissue loss is evident on MRIs.

It has been suggested that the right TP is the storehouse (Nakamura et al., 2000), or cite of recollection (Markowitsch, 1995) of personal, episodic memories. This hypothesis is based on findings showing that resection of the right TP can diminish that ability to recognize, or recall any information about famous or personally familiar faces (Fukatsu et al., 1999; Tippett et al., 2000; Glosser et al., 2003; Tsukiura et al., 2003). This deficit may reflect in part, a difficulty in recalling personal memories relevant to the test faces. A unique patient, RFR, after suffering damage to mostly the right and to a minor extent the left anterior temporal lobe, could generate names and information about family friends, but was unable to give any information about personal interactions or episodes pertaining to the same family friends (McCarthy and Warrington, 1992).

In contrast, the left anterior temporal lobe is more closely associated with semantic memory. Left resection surgery leaves the ability to generate information about people intact, but causes proper naming abilities (Glosser et al., 2003) and face–name associative learning (Tsukiura et al., 2003) to plummet. Similarly, left anterior temporal lobe atrophy in tv-FTD is largely associated with semantic memory impairments (Snowden et al., 2004). A review of retrograde amnesia cases found that left frontal/anterior temporal lobe damage usually led to loss of previously learned semantic memory, but left episodic memory intact (Markowitsch, 1995).

These findings raise the possibility that the right TP functions to link high-level sensory representations with emotional responses and social memory. The left TP functions to link high-level sensory representations, such as a face, with semantic information. However, more research is needed to clarify whether such a dissociation truly exists.

Social or emotional processing?

Any disruption of emotional processing will cause changes in social behaviour because normal social behaviour is predicated on emotion. Do the social deficits that are associated with TP damage go beyond emotional dysfunction? It is plausible that some of the social deficits reported in monkeys with bilateral TP lesions are actually due to a decoupling of emotions with perception and behaviour. For instance, the finding that mother monkeys forcibly reject their infants after TP damage may be due to the fact that the sight, smell and sound of their infant no longer arouses protective, nurturing maternal emotions, so the infant is rejected. Likewise, the failure to make and respond appropriately to social signals may be due to the fact that social interactions of any sort fail to elicit positive, rewarding emotions after TP damage. It is even possible that learned associations between perception and emotion are lost after TP damage, leading to inappropriate behaviour toward other monkeys.

Many of the strange social behaviours associated with tv-FTD can also be viewed as a decoupling between perception and emotion. It has been reported that such patients often become introverted and cold (Rankin et al., 2003), possibly reflecting a failure to derive pleasure or reward from social interactions. And as mentioned earlier, theory of mind tasks may rely on emotional processing to derive a social judgement.

However, there are some findings that do not neatly fit into the emotion category. First, there is a great deal of evidence pointing towards a role of the right TP in face memory and autobiographical memory, while the left TP has some role in personal semantic memory. These types of memory can be construed as types of social memory, and cannot be easily construed as fundamentally emotion-based. Second, the reviewed findings on neuroimaging studies of theory of mind indicate that TP activations are most frequently seen with tasks and stimuli in which there is some sort of social narrative, while simpler emotional stimuli rarely evoke TP activations. In sum, the data clearly indicate that the TP is involved in emotional processing while newer findings indicate that it may also have some role in social processing that goes beyond emotional processing.

Summary of social and emotional findings

The findings reviewed here support the functional role of the TP as a paralimbic region. The monkey lesion literature showed that bilateral TP lesions cause abnormal social and emotional processing. In humans, atrophy of the right TP and surrounding regions, found in right temporal variant of frontal temporal dementia, is distinguished by profound changes to personality, emotional regulation and social behaviour that mimic the deficits observed in monkeys with TP lesions. Other findings indicate that the TP is critical for aspects of face processing, specifically, recognition from auditory or visual cues. The former process may be linked to ventral aspects, the latter to dorsal aspects. Last, the neuroimaging literature links the TP to emotional processing of auditory, olfactory and visual stimuli and to mentalizing, or theory of mind.

We propose that a general function of the TP is to couple emotional responses to highly processed sensory stimuli. The mnemonic functions of this region allow for storage of perception–emotion linkages, forming the basis of personal semantic memory.

This hypothesis acknowledges that the TP is involved in high-level recognition as the end point of the auditory and visual ‘what’ streams, but also proposes that the functionality of this region extends beyond linking perceptual and mnemonic representations for recognition to include linkages with visceral emotional responses. Our review of the literature suggests that sensory streams are relatively separate in the human TP, suggesting that sensory-limbic binding is not integrated across sensory modality in this region.

Unfortunately, the paucity of systematic investigations of this region leaves many questions unanswered. For instance, we showed that there is a dorsal/ventral segregation of auditory and visual processing in the anterior temporal lobe. Is there any anatomical segregation of different types of emotional responses in this region? Also, a number of studies have found that ablations of the monkey orbital frontal cortex, the TP, and the amygdala cause a similar constellation of socioemotional deficits (Table 1). How do the roles of these regions differ in regards to social and emotional processing? There is evidence that the orbital frontal cortex and TP share many cytoarchitectural and response characteristics, as well as a close geographic relationship with neighbouring limbic structures (Kling and Steklis, 1976). Several findings, point towards a role for the anterior temporal lobes in emotional stability but more systematic studies are needed to flesh out its particular function. Emotion researchers have paid little attention to this region, with the bulk of interest centred on the nearby amygdala. We hope that the evidence reviewed here will spark future investigations of the social and emotional functions of this region.


We would like to thank the laboratories of Anjan Chatterjee and Geoff Aguirre for helpful discussion. Funding to pay the open access charges provided by startup-funds from the University of Pennsylvania to Ingrid Olson.

Appendix A

To select the studies used to create Fig. 2A, we prepared a list of published fMRI and PET studies found through an electronic search within psycinfo and pubmed databases. Search terms were: temporal pole, ba 38, anterior temporal lobe, planum polare, area TG, perirhinal cortex or paralimbic cortex. Then, we excluded publications that: (1) were not peer-reviewed, (2) studied special populations such as elderly or abnormal populations, (3) did not mention Talairach or MNI coordinates or (4) did not present material auditorily or visually. Studies that were included in the final list varied with regard to the type of stimulus and the type of test. These studies were: Sergent et al., 1992; Damasio et al., 1996; Imaizumi et al., 1997; Lane et al., 1997a, b; Gorno-Tempini et al., 1998; Griffiths et al., 1998; Phillips et al., 1998; Blair et al., 1999; Kimbrell et al., 1999; Dougherty et al., 1999; Brunet et al., 2000; Castelli et al., 2000; Dolan et al., 2000; Gallagher et al., 2000; Leveroni et al., 2000; Nakamura et al., 2000; Royet et al., 2000; Shin et al., 2000; Beauregard et al., 2001; Farrow et al., 2001; Gorno-Tempini and Price, 2001; Grabowski et al., 2001; Zatorre and Belin, 2001; Vaina et al., 2001; Berthoz et al., 2002; Koelsch et al., 2002; Lorberbaum et al., 2002; Moll et al., 2002b; Tsukiura et al., 2002; Eugene et al., 2003; Heekeren et al., 2003; Levesque et al., 2003; Schultz et al., 2003; Tsukiura et al., 2003; Warren and Griffiths, 2003; Brown et al., 2004; Grezes et al., 2004; Ranote et al., 2004; Tyler et al., 2004; Kim et al., 2005; Moss et al., 2005; Parsons et al., 2005; Platel, 2005; Pourtois et al., 2005; Rothstein et al., 2005; Satoh et al., 2006; and Tsukiura et al., 2006.

Appendix B

Studies used to create Fig. 2B were a subset of those used in Fig. 2A that manipulated emotion or theory of mind. These studies were: Fletcher et al., 1995; Lane et al., 1997a, b; Phillips et al., 1998; Blair et al., 1999; Dougherty et al., 1999; Kimbrell et al., 1999; Brunet et al., 2000; Castelli et al., 2000; Dolan et al., 2000; Gallagher et al., 2000; Royet et al., 2000; Shin et al., 2000; Beauregard et al., 2001; Farrow et al., 2001; Vaina et al., 2001; Vogeley et al., 2001; Berthoz et al., 2002; Lorberbaum et al., 2002; Moll et al., 2002a, b; Eugene et al., 2003; Heekeren et al., 2003; Levesque et al., 2003; Schultz et al., 2003; Tsukiura et al., 2003; Ranote et al., 2004; Grezes et al., 2004; and Kim et al., 2005.


  • Abbreviations:
    frontal temporal dementia
    temporal pole

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.


View Abstract